Forums :: Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives :: Books
Support This Site: Contributors :: Advertise


It is currently Dec Sun 16, 2018 2:28 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 82 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next
Author Message
 Post subject: Re: TV-2/U repair
PostPosted: Dec Wed 05, 2018 2:06 am 
Member

Joined: Dec Mon 05, 2011 1:10 am
Posts: 314
The offsets you have measured from the (presumed) "ideal" resistances corresponding to the scale readings should not have much impact on typical good vs. bad readings. BUT, if you are trying to measure "true" Gm, you probably want to dial in the corrected values for the three SHUNT settings that supposedly produce QUALITY meter readings directly scaled to Gm (37 for ranges B and C, 55 for range A and 86 for C, D and E). I haven´t seen the numbers behind these published (by Alan Douglas and others) settings, but I ASSUME that they are based on a 10x constant, linear relationship between the scale reading and the SHUNT resistance (i.e. 37 = 370 Ohms, etc.).


^^;;^^


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Wed 05, 2018 2:39 am 
Member

Joined: Mar Thu 29, 2018 3:41 am
Posts: 111
Location: Portland Oregon
I agree that the existing values are probably "close enough" for pass/fail, but I am trying to measure true Gm on this tester. As Alan Douglas stated, this tester has the potential to be close to the lab testers values.

I purchased Alan's book on tube testers and scanned most of it. I don't really understand his Hickok calibration procedure on page 6. I don't see where he has loaded his system and I have to believe that this will affect the readings. I believe that is how he came to the shunt values you noted. Can you help me develop the exact procedure for calibrating my tester using Alan's method? I have a variac that I've been using to set the voltage when I've been calibrating my tester. I imagine I can use this to provide the 50V source and then find the 10K Ohm, 50W resister (probably unobtanium).

I think that one can make a series of solid state Gm equivalent tubes like the example on Barry's website. This would be invaluable - a standard that won't change and is easily reproduced. I also have a buddy that has an Amplitrex 1000, but I'm not sure he ever has hooked it up to a PC to created a family of I-V curves. A family of curves for a tube and the Gm Values for the different tester conditions is needed for really good calibration standards.

I will definitely get the shunt offsets for the listed values. I have the shunt soldered in the tester right now, but I plan on taking it out tonight, removing all wires this time, and then remeasuring the linearity. I'll do a range of offsets and also measure the specific values listed by Alan (assuming his shunt pot was really linear).

I haven't touched the shunt pot in my tester. It seems to be pretty smooth. Would it hurt to put a little deoxit faderloob in it and work it back and forth? Do you think that could help improve the linearity?


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Wed 05, 2018 8:54 pm 
Member

Joined: Dec Mon 05, 2011 1:10 am
Posts: 314
I have used that procedure with a Variac, isolation transformer and 10 kOhms resistor to calibrate a number of TV-7 tube testers. I have never used it on any other models, and I don´t know what, if any, other models it may apply to. The procedure calls for a 10 k resistor with a fairly high power rating, but a General Radio decade box works fine ... just be sure to use only the lowest decade that gives 10 k (highest current capacity rating) and don´t run the current through it any longer than necessary (IIRC, the rated capacity is around 23 mA, or 5.3 W at 10,000 Ohms). I have always followed the Army TV-7 manual procedure to the letter ... not sure how that may differ from the one in the Alan Douglas book.

I once tried Deoxit on a TV-7 bias pot that had a "bad" spot (looked like poor contact at one specific point). Didn´t help. I think those military spec pots are pretty tightly sealed, so it´s probably tough to get the cleaner / lube to penetrate in to where it might help. I sure wouldn´t try it on any pot that didn´t seem to have a problem, like erratic, non-repeatable readings. I doubt it would have any effect on "accuracy" or linearity. As long as your pot is smooth and repeatable, that´s about as good as it gets.

I know absolutely nothing about making / using solid state "bogey" tubes, but I am real skeptical about the whole idea.



^^;;^^


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Wed 05, 2018 9:35 pm 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 2592
Location: Sunbury, Ohio 43074
I would not recommend spraying anything into partly or mostly sealed pots. Remember that what goes in, must be able to get out. As far as wafer switches, I'd opt for Deoxit on a Q tip directly applied, or use the spray/alcohol clean/air dry/bake method already discussed.

If there is no evidence of noisy or dirty pots or contacts, LEAVE them ALONE :)

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Thu 06, 2018 1:45 am 
Member

Joined: Mar Thu 29, 2018 3:41 am
Posts: 111
Location: Portland Oregon
Thanks for the reply guys. I'll leave my shunt pot alone.

I don't have the second 1:1 transformer and resistor listed in Alan Douglas' book, page 6, to do the calibration. I'll see if I can get one from the club and or order a wire wound, 10W/10kOhm 1% resistor. Is there any other way to validate proper Gm measurement on this tester? I don't know of anyone who has a calibration tube. I guess I could set my tester up to test like a TV-7, 539, etc. and then use the TV7 or 539 calibration tubes. I understand that the TV-7/539 has a 150V plate and a 130V screen (which are easy to set up on the TV2). I'll have to determine the grid voltage and the injection voltage. I expect that most of my testing will be at higher plate and screen voltages and maybe lower grid voltages (since my injection voltage is 0.25V). It would be great to qualify the tester in it's typical test conditions.

Do you think it would be worth while ordering a new shunt pot. I believe that the pot is a 1000 Ohm, 300 degree pot. Would this pot work?

https://www.nidec-copal-electronics.com ... jp-30b.pdf

I didn't get into the shop last night to remove and remeasure my pot.

There is a guy at Merren Audio that discusses changing the grounding scheme on the meters as well as changing the bias circuitry to a solid state DC and using the TV2 signal injection. I have done some precision electrical and understand the best practice of having a star ground strategy. I'll have to make some scope readings and see what the bias signal really looks like on my TV-2/U.

http://merrenaudio.com/tube_testers

I'll also measure the voltage drop and see how close the socket voltage is to the meter when the tester is under load.

The guy at merrenaudio also had a really interesting idea changing out some resistors with fuses. Apparently, there are two 47 ohm resistors in the circuit (I will get the exact resistors) that were originally used as fuses. The military did not allow in-circuit fuses and the BOM would not allow that the fuses be brought to the panel. The answer was to put 47 Ohm resistors in the circuit and if there was a short, the fuses would burn up. I don't know if a resister would be a very good fuse, but I believe that most of the resistors in the tester were 1/4W carbon resistors and I replaced a lot of them with 1/2 watt metal film resistors. I'm supposed to be getting the exact resistor numbers from the guy at Merrenaudio and will post them on this thread.

I know that this has been discussed in depth in the past, but what do people think about changing out the 83 and 6x4's with a solid state rectifier with the appropriate resistor as shown in Allan's book for the TV-2. I understand that the primary side of the transformers were a bit light on these and it may help reliability if one could lower the power draw in the secondary.

Dave


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Thu 06, 2018 8:29 am 
Member

Joined: Aug Wed 31, 2011 11:23 pm
Posts: 452
Location: Tucson, AZ
Not sure where you want to go with all these changes/mods. If you want to qualify a tester for Gm at different ranges, then it is very easy to make a calibration tube in a jig or on something like a uTracer and build a table of curves for different operating voltage and bias points. You then can measure the actual voltages on your tube tester and compare the results to the bogey tube. I typically setup a small signal tube like a 12AX7, 6SN7 and a pentode. There are some dual element tubes that each side can have their own set of calibration (Gm) values, I was sent one of these for my AVO tube testers. I have multiple tube testers and these measure within ~5% between the testers for a given operating voltage. The issue with using a tube calibrator, is that different versions of Hickok's have different adjustments and settings which will yield different readings. But if you measure the voltages under test you can then compare them to the bogey tube for those set parameters. This will give you a ballpark reading, these were never made to be precision testers. Vast majority of Hickok's are only good for weeding out bad tubes, the Gm results are pretty meaningless vs. running plate curves which look at multiple operating points.

Running filtered bias in a Hickok will give you erroneous readings. I have a 580A, which similar to the WE Hickok versions which use filtered bias and they readings are not comparable for the same test voltages in other Hickok's nor standard bogey tubes. One reason why the RD/WE versions of the 539 used different roll charts. This has to do with the nature of the plate and screen voltage being unfiltered DC. This has been outlined in previous postings. My 580A is modified for both filtered and unfiltered bias.

The newer variants on the 83 SS work just fine, I posted some of the earlier version using a zener dropping diode (applicable to fixed voltage single Hickok's) , used them in various Hickoks without any issues. That design has pretty much replaced everything else that people used previously. I have yet to see them not work in a Hickok tester. But if you have a working 83 tube, that also will work just fine. Diodes replaced the 83 rectifier in later Hickok's like the 580 series. The issue with fixed voltage Hickok's was getting the line set and the plate voltages to be correct, the Zener drops the voltage, plus the resistors. Some people add a fuse or just use 2 rectifiers followed by the dropping zeners. This has all been posted. I do not see that there would be an issue with the TV-2 since the voltages are all adjustable.

Grounding scheme, have not seen this as an issue with Hickok's. Ferrite beads were used on the tube socket wires in later models of Hickok's to attenuate oscillation seen with high Gm tubes. Star grounds and such are more an issue with low level signals and high frequencies, neither which should be factors in these tube testers. I ground the the tester's deck as a safety issue.


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Thu 06, 2018 1:59 pm 
Member

Joined: Dec Mon 05, 2011 1:10 am
Posts: 314
mksj wrote:
Not sure where you want to go with all these changes/mods. If you want to qualify a tester for Gm at different ranges, then it is very easy to make a calibration tube in a jig or on something like a uTracer and build a table of curves for different operating voltage and bias points. You then can measure the actual voltages on your tube tester and compare the results to the bogey tube.
> ...
Vast majority of Hickok's are only good for weeding out bad tubes, the Gm results are pretty meaningless vs. running plate curves which look at multiple operating points.

Running filtered bias in a Hickok will give you erroneous readings.
> ...
The newer variants on the 83 SS work just fine, > ...



Agreed, the "Mutual Conductance" readings of ANY tester based on the Hickok "Genius Circuit" are highly suspect and are not generally reproduceable between different models. There was a fairly long thread on here a while ago about whether the Hickok Gm scales represented outright FRAUD. IIRC, the consensus was that these testers all measure "quality" RELATED to, but not EQUAL to mutual conductance, and that Hickok should never have put those Gm scales on their meters in the first place.

Agreed also that if you really care all that much about getting a true, per the mathematical definition, Gm reading, then you need a setup designed to MEASURE that value, not some non-linearly related "genius" number. The uTracer or other curve tracer, or a simple (though tedious to use) test jig should do the job. Another way is to use a simple null-balance audio signal gain bridge like any of several General Radio models.

There have also been at least a couple of threads about using solid-state replacements for the #83 tube in Hickok circuit testers. Again IIRC, seems the consensus is that a well designed solid state replacement is actually BETTER than most #83 tubes, since the latter tend to be more "asymmetrical" between the two sides of the tube (somebody actually showed some really asymmetric #83 curves).

I´m surprised to read that using filtered DC bias and screen voltages are BAD. Seems I recall where the gurus mostly rate the Cardmatic, Bit-O-Matic, TV-2 etc. higher than others BECAUSE they run filtered, regulated voltages.???


^^;;^^


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Thu 06, 2018 2:45 pm 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 2592
Location: Sunbury, Ohio 43074
I suppose it bears repeating, in summary fashion, a few tru-isms about tubes and tube testers. (all of this available on this forum in older threads) Hopefully the OP can be dissuaded from devolving a nice TV-2 into an unreliable instrument.

Tube Testers were never designed as precision instruments but rather a go/no go tester despite the claims of their advertising. Some were better than others The methods by which they were produced at the factories is lost to time; what that means is all we can do now is try and reverse engineer, and mostly guess, as to what they did to make them "accurate". "Accurate" meaning ... as close an approximation as they could get to what an ideal tube should "test like" when measured on their tester.

Keep in mind that tubes as manufactured had, and still have, a range of acceptable Gm, noise, gas, name your parameter. The tester was designed to weed out good from bad, and weak from strong .... NOT as a precision measure of any parameter of that tube. Even if you "MATCH" two tubes for Gm using the tester, you have only matched it at ONE specific operating point within the tube's characteristic curve(s).

As far as calibration, the methods used by any manufacturer to make the so-called "bogey" tubes is lost to time, although one might surmise that they were produced using a method similar to what mksj refers to..... a very accurate setup to actually measure performance after, presumably, burning in a tube for (insert your guess here) hours, then hand picking the best, and THEN subjecting them to this analysis. And you STILL end up with a tube that's only good for one particular make/model of tester, although presumable with a 'known' Gm figure.

My guess is that the roll chart values for tube testers were created FOR EACH MODEL after testing dozens of each of a range of selected tube types, and then generating the chart values based on THAT tester model's design criteria and performance. I'm sure they also consulted the tube manufacturer specs as well during this process.

What you end up with is a non precision device (the tube tester) based on a non precision component (the tube), that has been optimized to test said tube at one single operating point in it's characteristic (by design) Gm curve.

There is a rapidly diminishing point of return in trying to make a tube tester "more accurate". In re-designing one, as the OP seems to be attempting, I fear that the end results will be a rather less accurate instrument than the original design called for.

The best one actually needs, and perhaps should even hope for, would be to restore these instruments to their original factory specs as much as possible, and stop there. They worked as intended back in the day, and they still do, and they still will, assuming only that they are in good repair and calibration. And there's the rub. We're back to needing a calibration tube :)

I'll let former forum threads stand as far as discussing this.

You want accuracy and exact results? The CURVE TRACER has been mentioned a number of times. The uTracer costs far less than what the OP seems to be spending on this "upgrade", is far more accurate, and reads out the entire operating curve of a given tube. And that's but one of the available curve tracers.

color me confused..... why all the effort to make something into what it never was and never will be?

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Thu 06, 2018 11:36 pm 
Member

Joined: Aug Wed 31, 2011 11:23 pm
Posts: 452
Location: Tucson, AZ
Morcegao wrote:

I´m surprised to read that using filtered DC bias and screen voltages are BAD. Seems I recall where the gurus mostly rate the Cardmatic, Bit-O-Matic, TV-2 etc. higher than others BECAUSE they run filtered, regulated voltages.???
[/quote]

Regulated supplies are good, cardomatic uses regulated power supplies so the bias also needs to be regulated. The TV-2 was all unfiltered power supplies, but the voltages are adjustable under load. The problem is with the RD/WE versions of the 539 and the 580 models is both the screen and plate voltages are unregulated (unfiltered DC) but the bias is regulated (filtered). When the bias is unfiltered the waves all align so the voltages are all an averaging of the different waves are in sync., with filtered bias this is not the case. This was outlined by Cdoose in an old thread "A tale of two tube testers". Below are two graphs which he shows that the plate current and Gm are inaccurate with this combination. I also noted this on my 580A and couldn't figure out why the readings never matched other testers results with tubes being tested at the same specs. I subsequently pulled the wiring for the bias to the top deck so I can switch between unfiltered and filtered bias, along with a balance pot for the bridge balance adjustment for the different plate taps.


Attachments:
6SN7 Transfer Curves.jpg
6SN7 Transfer Curves.jpg [ 106.24 KiB | Viewed 112 times ]
6SN7 Plate current filtered vs unfiltered bias.jpg
6SN7 Plate current filtered vs unfiltered bias.jpg [ 104.81 KiB | Viewed 112 times ]
Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 12:19 am 
Member

Joined: Mar Thu 29, 2018 3:41 am
Posts: 111
Location: Portland Oregon
Barry. I get what you are saying about making a tube tester a pile of poop by changing too many items with a good chance it won't work. I am simply floating ideas to see what people say. I haven't done any significant changes to my tester except add new caps, change their values slightly, replace out of spec carbon resisters and change the values so they pass the echelon 5 tests better. The tester seems to be much better than what I started with and I just want to pursue making it as good as new (or maybe a bit better if I can make the change w/o compromising the tool). One of my remaining questions if I should just set up a calibration with the existing shunt pot or try to replace it. I"m inclined to set up a table to calibrate the shunt pot and then wait for one to come up for auction and then improve it. I am also wondering how people eventually qualify the tester if they do not have access to tubes that were created on a curve tracer. I currently don't have a power supply that provides good ranges of plate and grid voltages to make a test jig. I would like to get a reproducible method to create a Gm for the different ranges on my tester. I would actually prefer something solid state.

The TV2 has a capacitor on the grid bias circuit (C3, which I increased to 100uF) that looks to be some sort of filtering function. It may be better to have a couple of capacitors on the circuit but I haven't measured the wave form with my Oscilloscope. I understood (agree with Morcegao) that this filtering was one of the items that made the TV-2/U superior to other testers, along with the ability to independently change the tube under test voltages and the reasonably low injection signal. I read that if one is very careful, and made the adjustments with a tester that had all of the components close to nominal (I'm using 1% resistors), that one can get the TV2 to correlate withing around 5% with lab testers. This is similar to what Allan Douglas measured in his book (and seems to be in agreement with mksj statement - unless I am reading his post incorrectly). This is what I'm trying to achieve: a tool that can be shown to be correct enough for a number of measures to insure tube health (which includes a measurement of Gm or potentially a good enough measure of Gm). I'll have to find that thread that Morcegao referenced which focused on the definition of Gm and what the Hickok circuit is testing. At first blush, it seems to be a measure of Gm (dIp/dVg), but I'm interested to read what others think.

I don't have a curve tracer but I agree, this is the best way to measure tube performance. I know a guy who has an amplex1000, but I'm not sure he knows how to hook up the PC and make IV curves. I'll ask. I looked at the uTracer really close but I think I will eventually get the eTracer. The eTracer seems like a much more capable tool and the guy is still really active in improving the software, etc. It also comes with a box, sockets, etc. and uses jumper wires as opposed to the rotary switches (a simpler, reproducible albeit slower to set up design).

I will also like to find the design for a well designed solid state 83. I would think that this may be a much more stable, matched source but most importantly, it will reduce the power on the primary transformer T2 coils. From what I can see, the TV-2/U T2 primary side has a tendency to burn out and it may be better to replace the 83 with a well designed 83ss. Does anyone have a similar design for the 6X4 tube?

I agree that the star ground is very important for low level signals (I was measuring pA and fA) which is ~ 6 orders or magnitude less than what we are measuring here. I was told that all of the grounds were connected together. I was thinking that a star ground would help isolate the meters better and reduce the parasitic wire resistance.

I am looking for the resistors that the guy from Merran Audio was referring. I will contact him directly and get the resistor numbers. I don't know if R52 is one or not...


Last edited by TechyMechy on Dec Fri 07, 2018 6:31 am, edited 1 time in total.

Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 2:30 am 
Member

Joined: Aug Wed 31, 2011 11:23 pm
Posts: 452
Location: Tucson, AZ
If you would like, you can mail me some tubes that you would like tested and I will run them on my uTracer, AVO CT-160 and/or 580A at set voltages or run plate curves. I typically setup the tubes to run a published operating values for tube testers that have those capabilities. You can send me an email address through the PM in this forum if interested. Take me about a week to get everything out and tested, include a return label with postage to get them back to you. I will say that the AVO and uTracer are very close in measured values, the 580A is not far behind but it has been modified to correct known design flaws. I also have a 539B with SS 83 and 5Y3, and when compared at the same operating parameters on the uTracer, it is also within 5% of Gm and Plate current. All my tube testers have digital readouts for all the operating voltages and plate current, so there is minimal error in reading the values.

The uTracer is inexpensive and surprisingly accurate within it's operating limitations. I use an external filament supply which works much better than the onboard one. The eTracer looks very interesting and more capable but at 4X the cost of the uTracer, can't really see upgrading when I get such good results with the uTracer. I have mine setup with AVO type thumb wheels and wireless computer connection so very easy to operate.

The solid state 83 designs posted by me and subsequently by Mike Higgin's work just fine, they model very closely to the 83 tube and the same design has been used in countless Hickok testers. This is a series of videos with older SS 83 designs, all did not work until part #3 which uses the zeners. Cost is a few dollars in parts. The only thing recommended by others was using two diodes per leg instead one one so if it fails closed your transformer won't be smoked. That being said, I used 3A 800V diodes, I can't see them ever failing at 150V and under 100 mA load. As far as other rectifier tubes and SS, each one needs to be modeled and a suitable design developed. I have a design for the 5Y3 which works fine, but these tube last a long time and I would not try to fix what isn't broken unless you have some way to validate the results.
https://www.youtube.com/watch?v=j47uN1kjnCw

On a lot of other mods, I would be a bit more leery of making changes without a solid rational and a way to verify the results. The Hickok's work as designed, there were lots of compromises/changes and individual tester mods to get them to read/test a particular way. Some Hickok's you adjust voltages under test, others you do not and the voltage sag is part of the reading for that particular model/roll chart. The TV-2 is no exception in that the plate is unfiltered, the screen is filtered with a 10uF capacitor and the bias is filtered with a 50uF capacitor. The regulation is based on voltage dividers and adjustments based on meter readings. My experience with the 580A (as I tried filtering the plate, screen and bias) was mostly with negative results when comparing the before and after changes to known tube values. I use a 6SN7 as my default bogey tube, mostly because of its stability and wide range of plate and bias voltages that fit Hickok testers. The repeat deviation on retesting is negligible. Power tubes, I use either a 6L6 or EL34, but smaller envelope pentodes have less drift with repeat testing.


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 3:58 am 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 2592
Location: Sunbury, Ohio 43074
One of my remaining questions if I should just set up a calibration with the existing shunt pot or try to replace it. What makes you believe there is something wrong with your existing shunt pot? Maybe I missed some earlier post....

I"m inclined to set up a table to calibrate the shunt pot and then wait for one to come up for auction and then improve it. How are you going to verify that some calibration table is correct? These testers were engineered to refer to a specific set of values on a roll chart (or cards or whatever) and if you "make it better" you risk throwing those posted values out the window

I am also wondering how people eventually qualify the tester if they do not have access to tubes that were created on a curve tracer. Again, you are striving for precision, where none ever existed and where none can really exist, nor is there such a need. I'd define "qualify" to mean "is in calibration according to the setup procedures

I read that if one is very careful, and made the adjustments with a tester that had all of the components close to nominal (I'm using 1% resistors), that one can get the TV2 to correlate withing around 5% with lab testers. I hope you also read that in some cases, Hickok hand selected resistors during the manufacturing process. I don't know that is the case in the TV-2, but I'd assume it to be so. IF yoiur resistors are within spec as to what they SAY on them (not what the parts list says), replacing them is going to yield little if any better results, and may be harmful to calibration in some way

There is absolutely no reason to re-engineer the ground system in a tube tester. Just make sure the existing system is fully functional (no corrosion, loose connections, etc). Star grounds vs other .... not going to make the slightest bit of difference until you get way up there in frequency, or are running very long cables in, say, a studio.

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 6:15 am 
Member

Joined: Mar Thu 29, 2018 3:41 am
Posts: 111
Location: Portland Oregon
Attachment:
Shunt_12_6_18.JPG
Shunt_12_6_18.JPG [ 25.93 KiB | Viewed 87 times ]
Hi Barry,

Thanks for your questions and thoughts.

Attached is the actual data for my shunt pot. I know that this is supposed to be a 1000 Ohm pot and I'm pretty sure it's supposed to be linear. As you can see, the errors become pretty big above the 50 setting.

I am assuming that the pot is supposed to be 1000 Ohms and linear. What I have done is take a resistance reading at each 10 units and add/subtract units to achieve the theoretical shunt resistance. As you can see, some of those offsets are as large as 5 units. Then I was going to do a linear interpolation between each 10 units to provide an estimate for the shunt setting/resistance value between each 10 units. It's pretty easy to do this in excel to provide an offset table. Of course, I'd get closer if I do it every 5 units (20 readings) or 1 unit (100 readings).

As far as calibration, I couldn't complete the Shunt Control test in the Echelon 5 tests. I don't have the special instrument to complete the test. I figured that I could measure my meter and my shunt pot and then test some tubes. It would be great to have known Gm tubes for certain settings so I can tell if my tester is providing correct results. Alan Douglas tested his tester and it was reasonably close to the results from his lab tester. I have no reason to believe that my tester couldn't be reasonably close to a lab tester or curve tracer results. The Echelon 5 testing does have testing tubes as part of the qualification procedure. I just don't have access to any known tubes. If the solid state tubes work, I'd be willing to build a couple of solid state Gm surrogates. I actually think that qualification would be best with known tubes plus a solid state device. If they all align, and if you do a small gage study, one should get a pretty good idea of the reproducibility and repeatability of your tester. This is pretty easy to do and it really give you an idea of what you can and cannot say about your tool. Why am I doing this...because it's fun.

As far as replacement of components, Hickok never ever made the TV-2. My tester was made by CBS Columbia (I don't think they outsourced to Hickok). The manufacturers use the Hickok circuit. The manual does not have the component values in it but the schematic does. I've been really careful to replace the resistor as was found except for a couple of resistance values that when changed, aligned the meters with the voltages better. I've been replacing the carbon resistors with 1% metal film. I believe that most of the resistors are 1/4 W (est from size), but I've been replacing them with 1/2W metal film.

I agree with your assessment about the star ground.

mksj,

When you mentioned having two diodes per leg, I am assuming that the diodes are in series so that if one locks closed, the other is working - right? I might build a SS83 just to reduce the power on my transformer. I've read more than once that the T2 transformer primary side is one of the items that fails and kills a tester.

I will have to more carefully read the operators manual for the TV-2, but I believe that the manual I have says to adjust the voltages after the tube is under test. If you use a test socket, you can dial in the voltages very closely. That's one of the nice parts of the TV-2. The screen has two 10 uF capacitors, I increased the size of my bias capacitor to 100 uF and I have a 2200 uF capacitor for C1. I cannot figure out what C1 is used for in the circuit..it goes to post 11 on switch 2, section 2, front but that is where the schematic ends. There are so many errors in the manual. I think there are errors in the schematic too but you have to look at these a lot to make sure you have the switches right, etc. too.

I'll send you an IM and let's exchange emails. That is so nice of you to offer to measure some tubes for me. I'll ping the club for a 6L6 and a 6SN7. I saw in a previous post that you made the calibration 6SN7 solid state device. Are you still using this and recommending it? I move really slow. Between work, family committments, home projects, chores, etc....there is very little time for this hobby right now. I expect it will take me 2+ weeks to find these tubes. I have zero stock.

You should post about your uTracer. I was seriously thinking about buying one and assembling it, but by the time I had the power supply, switches, cabinet, etc, I would have about $500 into it. I think I will go with the eTracer because it has quite a bit more capability and seems to use more components (updated controller, etc.) The uTracer is an amazing device...and Ronald is a top shelf, highly talented/intellegent guy (I've exchanged emails with him).

As far as designing a 6x4 - all of these diode tubes look really similar. I fully don't understand why Alan chose two 10Ohm resistors on filament legs when the tube has a 5V supply and a 3A filament current (according to the spec). I am also amazed that the 5.6V zener and diode model the electron flow (max 800 mA plate current) from the plate to filament. For the 5Y3G, the filament voltage is still 5V, the current is 2A but you use the same 10 Ohm resistors to model the filament resistance and then tie that to a 25Ohm resistors with 12V zener and a diode to model the electron flow (400 mA max plate current).

How the heck does a guy model a 6x4 which has a 6.3V, 0.6A filament with a max 245 mA plate current? The 6x4 tube has a cathode, the 83 and 5Y3G do not. The tubes are also not operated at the same plate voltage on V3, V2 has the cathode at 0V and both plates are the same voltage. Is there software for modeling this?


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 9:13 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 589
Location: Stafford, Texas USA
TechyMechy
Where did you get this?
“I am also amazed that the 5.6V zener and diode model the electron flow (max 800 mA plate current) from the plate to filament. For the 5Y3G, the filament voltage is still 5V, the current is 2A but you use the same 10 Ohm resistors to model the filament resistance and then tie that to a 25Ohm resistors with 12V zener and a diode to model the electron flow (400 mA max plate current). “
“How the heck does a guy model a 6x4 which has a 6.3V, 0.6A filament with a max 245 mA plate current? The 6x4 tube has a cathode; the 83 and 5Y3G do not. The tubes are also not operated at the same plate voltage on V3, V2 has the cathode at 0V and both plates are the same voltage. “

The two 10 ohm 1 watt resistor are added to the 83 and 5y3 to make an artificial center tap/cathode.

Since 6X4 has a cathode it does not need the two 10 ohm 1 watt resistor for the filament. All you need is two diode in series with two resistors Plates to cathode.
The value of the resistors need to be big enough to match the voltage drop of the 6X4 that 22 volts at 70 mA.

Go hear for the right way to build SS 83 and 5Y3.

http://www.byan-roper.org/m_higgins/201 ... ration.pdf
Jimmie


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 9:21 am 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 2592
Location: Sunbury, Ohio 43074
[quote="TechyMechy"]
Attachment:
Shunt_12_6_18.JPG
Hi Barry,

Thanks for your questions and thoughts.

Attached is the actual data for my shunt pot. I know that this is supposed to be a 1000 Ohm pot and I'm pretty sure it's supposed to be linear. As you can see, the errors become pretty big above the 50 setting. I am assuming that the pot is supposed to be 1000 Ohms and linear. That's an assumption based on what? Your readings do not agree with that theory. A pot can fail in several ways, but changing it's taper is not one of them. Changing facts to fit theory isn't a good idea in this case :)

As far as replacement of components, Hickok never ever made the TV-2. Hickok designed the tester and as with all military contracts of this type, at the time,, licensed it out to various manufacturers

I really don't understand the rush to replace the tubes with solid state devices. The transformers were designed to operate with the tube load, and as long as the units do not overheat, they should continue to do so just fine. If you are hell bent on eliminating the tubes, I'd highly suggest you get the unit 100% working and calibrated WITH the tubes first, and THEN fiddle with things one at a time. I have no intention of replacing the tubes in any of my Hickoks, probably ever. The "theory" (there's that word again) that they kill power transformers in these instruments has no basis in fact.... it's just hype. Can you reduce the load on a transformer? Yes, you can. But it also changes other parameters unless you exactly simulate the loads that the tube used to place on the transformer anyway. Seems pretty close to a circular equation to me, at least while these tubes are still readily available. Just my preference......

I'll say one thing about the uTracer .. it's a wonderful device for it's cost. I'm not sure where you intend to buy the support materials (switches, sockets etc) but it can be done for well under the price you mention. That said, nothing wrong with the Etracer either. I built a uTracer but haven't had the time to make a cabinet for it. One idea I had was to just repurpose an otherwise dud tube tester, but I'll probably just build from scratch.

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 1:12 pm 
Member

Joined: Dec Mon 05, 2011 1:10 am
Posts: 314
TechyMechy wrote:
Attachment:
Shunt_12_6_18.JPG

Attached is the actual data for my shunt pot. I know that this is supposed to be a 1000 Ohm pot and I'm pretty sure it's supposed to be linear. As you can see, the errors become pretty big above the 50 setting.



I believe that the resistance actually shunted across the meter is not what you measured, but actually the DIFFERENCE between the total resistance of the pot winding and what you measured. If that is true, the shunt value at 90 would be 1023 - 62 = 961 Ohms. The resistance at 10 would be 1023 - 914 = 109 Ohms.

My best guess (and that´s all it is, a GUESS) is that the ideal, "design" shunt resistance = 10 x the scale reading (in Ohms).

if you can figure out where to set the knob to get 370, 550 and 860 Ohms, you should have the proper settings to read Gm, per the table in the Alan Douglas book.


^^;;^^


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 7:30 pm 
Member

Joined: Mar Thu 29, 2018 3:41 am
Posts: 111
Location: Portland Oregon
Hi Folks,

Thanks for the discussion:

Jimmy, mksj gave the link for a proper designed substitute for the 83 mercury tube: https://www.youtube.com/watch?v=j47uN1kjnCw

The link for the ARF thread: viewtopic.php?f=8&t=181387&hilit=Updated+Hickok+539+Calibration

The numbers I quoted were off of the tube data sheets and manual (I used the max plate current from the data sheets, the plate voltages that were specified in my TV2 manual as well as the filament voltage and current on the data sheets). I honestly don't know how to model thermionic emission and how fields/space charge effects can be modeled via solid state devices. I thought that the zener diode and silicon diode simulated the field to get the current flowing and then the resistor provided the "filament to plate resistance".

I'm intrigued with your thoughts on the 6x4 tube. My manual says that the cathode on V2 is at ground (0 volts) and the two plates are operating at the same voltage (-165VDC with resistance to ground 8500 Ohms). The cathode on V3 is at 100Vdc, 12K Ohms to ground, Plate 1 is at 80Vac/80 ohms to ground, and plate is 210Vac/80 ohms to ground.
Will a simple diode (0.5 Volt drop) followed by a resistor ~307 Ohms (21.5V/.070A) going to pin 1 be right? I'm not too sure how to handle imbalanced operation on tubes. I would really need an IV curve. Plus, I need to read up on how to properly design a solid state replacement tube.

Morcegao, I think you are right as far as measuring the shunt pot. I haven't been able to get out in the shop this week. I'll get out, hopefully, tonight and remeasure the shunt pot to see how it behaves with no wires soldered to it. I don't know if the ideal shunted resistance is 10X the scale reading. If that is the case, my shunt is still not very good, but this should be reconfirmed with a separate reading with the shunt pot in the tester and with no wires attached.

I need to get a couple of tubes that have been fully characterized on a curve tracer. This will really help me determine the correct settings per Alan D's recommendations.

Barry - I'll be really careful with my change control. I won't be changing things w/o experimentally verifying the results. This could include the solid state 83 or anything. I am using this form to spark conversation, thought, and ideas of what could be going on with my TV2 and how to best fix it. There is NOBODY locally, who I know, has experience with the TV2.

Dave


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 7:36 pm 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 2592
Location: Sunbury, Ohio 43074
Dave, there really aren't many anywhere with experience in these... which is why I persist on cautioning you to get it working before you try to do any mods

..... I've not even had time to play with my own yet. It's about #6 on the list once I get back to the museum, which will commence with tube testers research.

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 10:57 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 589
Location: Stafford, Texas USA
TechyMechy
From your statement’s the first thing you need to do is get a book on Tube’s an read it.

http://www.nutshellhifi.com/library/MostVacuumTubes.pdf Getting the most out of Vacuum Tubes Written in 1960 by Robert B. Tomer, a CBS Electronics engineer, and an essential resource for all working with vacuum tubes today

http://www.tubebooks.org/ Here you will find a collection of vintage engineering texts, all of the RCA Receiving tube manuals, other obsolete information, presented free of charge and without annoying advertisements. DIY-audio-related information, visit www.pmillett.com
https://www.americanradiohistory.com/ history of radio and television in a library of thousands of magazines and publications

“I'm intrigued with your thoughts on the 6x4 tube. My manual says that the cathode on V2 is at ground (0 volts) and the two plates are operating at the same voltage (-165VDC with resistance to ground 8500 Ohms). The cathode on V3 is at 100Vdc, 12K Ohms to ground, Plate 1 is at 80Vac/80 ohms to ground, and plate is 210Vac/80 ohms to ground.”

V2 is the Bias supply that – neg. that come off Tap/pin 27 of the transformer about -150 DC going to C3
V3 is the Screen supply that + pos. coming off the Pin 7 of the 6X4 Cathode going to C4A this voltage
will change with the setting of S3 Plate and Screen range switch and pot R54, Plate 1 Pin 1 and Plate 2 Pin 6 will be set by S3 to the same voltage on both plates.

Learning points
1. Vacuum Tubes have a tolerance -20 to +40
2. There is no such thing as a “correct” (or single) mutual conductance score. Mutual conductance is a result of the operating point of the tube (plate voltage, signal voltage, grid bias, etc.) or for that matter emission is to.
3. Tube Tester Factory setup data OFTEN provides a substandard operating point for the tube in question. This is sometimes because of mistake or carelessness in creating the setup data, and sometimes due to design limitations of the test circuit (one fixed signal voltage that is substandard for a particular tube, or a fixed plate/screen voltage that is substandard for that tube).
4. All tube testers are a compromise in functionality, compromise in accuracy, and none are perfect. No models test for all tube characteristics. Not even the tube manufacturers themselves had test equipment that could detect all bad tubes, or verify all good tubes, or test for all tube characteristics. A 100% accurate tube tester is not even possible because a tube can work excellent in one circuit and not work at all in another circuit.
5. KNOW YOUR TUBE TESTER is of utmost importance, not blindly relying on the results that you see on the meter.
6. IN all RCA data books the final test is, will the tube work in the circuit that it being used in.
7. I have 40 years on test tube in the 60’s to test the latest Semiconductor in 2000’s.
8. Before I retired I had the best tube tester that money could buy a $1,000,000 plus high power Semiconductor tester, do miss it.


Top
 Profile  
 
 Post subject: Re: TV-2/U repair
PostPosted: Dec Fri 07, 2018 11:28 pm 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 2592
Location: Sunbury, Ohio 43074
Jimmie.......... (asume texas drawal) ... we'all been tryin' to tell him all that.........

well said :)

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
Post New Topic Post Reply  [ 82 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: Chas, Jim Beckford, Jim Mueller, Jimmie Stewart and 9 guests



Search for:
Jump to:  
























Privacy Policy :: Powered by phpBB