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


It is currently Dec Sun 17, 2017 12:56 pm


All times are UTC [ DST ]





Post New Topic Post Reply  [ 59 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 3:42 am 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
I'm not sure which was the first darlington mentioned. The one I said was massive overkill?.... and might not have enough gain at 200mA Ic

Based on your understanding of the circuit and the requirements, what gain do you think is required?

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 3:52 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
I meant the transistor I listed with the power transistor you said looked like it would work.


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 4:01 am 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
Tube Radio wrote:
Would it be possible to just substitute that power transistor for the regulator tubes?

How would that work? Aren't the tubes 2-terminal devices? (Effectively vacuum zeners)

With regard to the transistors: How much total gain is required?

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 10:43 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
Meant to say the two pass tubes which are triodes.


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 12:10 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
Tube Radio wrote:
Meant to say the two pass tubes which are triodes.

Well....a junction transistor needs base current.....I think a tube has to be replaced by a MOSFET.

Two ways to proceed here:
1. Look up the transistors in the original design so you can find the right modern replacements. Build according to the original design.
2. Start with a clean piece of paper and work up a new design. For this option, the first step is to define the requirements.

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 12:30 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
Requirements needed:
Output voltage and tolerance (not the same as regulation)
Supply voltage range
Maximum load current
Regulation with respect to:
--load current
--temperature
--supply voltage

To be rigorous, need ripple specs also, but I think the other requirements will dominate.

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 12:40 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
For now I will stick with the solid state circuit and try breadboarding it with the one transistor just to see how well it works.

If it doesn't work good enough I'll try it with the two transistors.

I may buy a second power transistor and try to substitute it for the pass tubes just to see if it will work.

As a reference here's the original regulator circuit.

Attachment:
Original regulator.jpg
Original regulator.jpg [ 82.23 KiB | Viewed 509 times ]


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 1:03 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
Do you know how the original circuit works?

Do you understand why you would need a power MOSFET to replace the pass tubes?

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 1:23 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
Just was a thought and seeing that it would require more work to use the original circuit I'll just go with the solid state circuit.

Now the solid state circuit as is according to the original author met if not exceeded the spec of the original R-390 regulator circuit.

It can be found here

http://www.navy-radio.com/manuals/hsn/h ... e%2052.pdf


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 1:34 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
So.....use the S/S circuit as designed.....find the the specs on the transistors and use the modern equivalents.

Looking at the original, I think it has better regulation than what I was guessing. When you get all done, do some testing to compare the performance between the original and the S/S.

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 2:19 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
Here's the data on the 2SD1398

http://www.datasheets360.com/pdf/-7516738506066350219

Here's the datasheet on the 2SC039

http://pdf.datasheetcatalog.com/datashe ... SC3039.pdf

Both are way overkill for this application, but the circuit the author found in his R-390 used those as it was most likely what the builder had on hand that would work.

I'll try the power transistor I listed https://www.mouser.com/Search/ProductDe ... 2-BD159STU

Will start with one and see how it works when breadboarded.

If one isn't enough I will add the second one in the darlington configuration.

Here's two updated schematics.

Attachment:
Regulator 3.jpg
Regulator 3.jpg [ 61.03 KiB | Viewed 504 times ]


Attachment:
Regulator 4.jpg
Regulator 4.jpg [ 53.95 KiB | Viewed 504 times ]


I got the parts ordered including the two diodes and first filter cap of the R-390 (10uF on the regulator schematic) so that I could breadboard the circuit independent of the R-390 so that I can get it working properly before installing in the R-390.

Now in order to test the regulator under load I will need a 1K 32.4 watt resistor which will provide a 180mA load at 180Vdc.

Depending on the heatsink needed I may be able to build the circuit on an octal tube base. That would allow for returning it back to original fairly easily by unplugging the circuit, reinstalling the 47 ohm resistors putting the power supply resistors back how they were stock and plugging in both regulator tubes and the other voltage reference tube.


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 3:51 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
Why would you use one power transistor to drive another of the same type? Remember that the gain of a typical power transistor will be lower if operated at a current much lower than it is designed for.

I think maybe small-signal darlingtons might be fine with identical transistors, but--in this case--I'd use a driver transistor optimized for the current it will be handling.

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 4:15 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
Here is a good article on the error in feedback amplifiers. This would be used to analyze load regulation.
https://web.stanford.edu/~boyd/ee102/fdbk-static.pdf
The line regulation is a bit more difficult. I think one would start by analyzing the stage gain to see how it varies with supply voltage.

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 4:27 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
pixellany wrote:
Why would you use one power transistor to drive another of the same type? Remember that the gain of a typical power transistor will be lower if operated at a current much lower than it is designed for.

I think maybe small-signal darlingtons might be fine with identical transistors, but--in this case--I'd use a driver transistor optimized for the current it will be handling.


The reason is because the article mentioned both transistors can be the same type.

In fact judging by the article it would appear that most any transistor that meets the voltage and current requirements would work.


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 4:50 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
No!!!!......the transistors have to also meet the GAIN requirements.

The factory circuit you showed is pretty sophisticated. I will not be surprised if you discover that it outperforms the S/S replacement.

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 6:02 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
From the linked article.

Quote:
Performance Comparison & Specs:
This brought up the need for some empirical testing. Barry had a stock R-391 handy to use as a
benchmark and put out a call for assistance on the R-390 e-mail reflector for other R-390/391
owners with a variac and voltmeter handy.
We varied the line voltage from 95 VAC up to 125 VAC in 5-volt steps and measured the B+ at
each, monitoring the AC voltage with a meter. The solid state circuit performed comparably to all
but one Motorola-built R-390 which, for some reason, far exceeded specs.
The Darlington circuit showed a delta of +0.4 -0.6 from the 110 volt input value. The original R390
design specification was+/- 1 volt. The solid state regulator did a delta of 1.0 volts. My second zener
mod will make the solid state version do better.



Here's the specs of the original regulator circuit in the R-390 from the same article

Quote:
Input voltage variation of ±15% B+ variation ±1 volt maximum
Load current variation 100 to 200 ma B+ variation ±2 volt maximum
Ripple voltage normal 0.01 volts
Ripple voltage worst condition 0.1 volts"
So, the simple zener/Darlington supply basically meets the R-390 engineering specifications.



Transistor types from the same article.

Quote:
Thus, the needs of Q 1 and Q2 can be met by most any transistor that can handle a few amps and
more than 350 volts in a T0-220 or isolated package. Both can be the same type. The original
2SD1398 included a 50 ohm resistor from base to emitter and a clamp diode from collector to
emitter. The resistor would be a bother in Q 1 causing poorer regulation. If not present in Q2, a
resistor (R4 on the schematic) with value between 500 ohms and 2K should be there to divert
leakage from the input of Q2. In this circuit, the clamp diode is of no benefit, but it won't hurt
anything either.


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 6:09 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 20683
Location: Annapolis, MD
OK!!--build the circuit as published

_________________
-Mark http://pixellany.com

"It's always something". --Gilda Radner (1946 - 1989)


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Tue 24, 2017 6:31 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
The plan is to build as published since it works according t the author and see if any refinements need to be made.

I still like your idea of possibly only needing one transistor so I will test that first.

If it doesn't work as it should I'll go with two transistors.

If the transistor I selected does not work adequately then I'll definitely study the circuit more carefully and continue seeking advice.

Also thanks for all the help you have provided.


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Sat 28, 2017 1:14 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
Good news.

I breadboarded the circuit using the same rectifier and filter cap setup as in the R-390 minus of course the 47 ohm resistors in the power supply right off the diodes as I cannot remember how the resistors are connected given I replaced the rectifiers with diodes when one heater failed.

I tried it with one transistor.

The preliminary numbers are in.

251Vdc in

181.3Vdc out

50mA load current.

I still gotta figure a way to get 200mA of current

The 1K is easy to find at work. Finding it in 32.4 watt capability, however isn't.

That would give me 180mA which is close enough for testing.

If it does work good with one transistor at 180mA of current then I'm all set. I got the transistor mounted to a strip of metal like one would find as a divider in a parts drawer as a makeshift heatsink.

If the transistor doesn't get hot with just the strip of metal I may be able to mount it directly to the divider on the audio chassis between the regulator tubes and other tubes.

If this circuit does indeed work good enough it may be something I send to one of the people with a R-390 page given this circuit is relatively simple and uses currently available parts.



EDIT:

Nope will need two transistors.

One wasn't enough.

Got it going with one then tried to load it with nearly 200mA.

Heatsink wasn't good enough for the transistor and it popped and in the process my 180 volt zener shorted.

While trying it with the other transistor and the 180 volt zener disconnected apparently the base current is too great for the resistors in the circuit as I couldn't get any higher than about 124Vdc.

That transistor I currently have may work for a drive transistor though.

EDIT 2:

I do have a 2N6770 MOSFET that I can try.

https://www.microsemi.com/document-port ... s-0101-pdf

Can that directly replace the pass tubes?

If so is it drain to where the plate connects, gate to where the grid connects and source to where the cathode connects?

Provided it will work I can use an octal socket to connect the MOSFET and find a way to secure the heatsink. That way the mod is easily reversible by simply unplugging the MOSFET then installing the pass tubes. Plus it will regulate better given it has feedback.

I also have a 2N3902 which will work in the circuit I posted.

https://cdn.macom.com/datasheets/2N3902_2N5157.pdf


EDIT 3:

Tried the power mosfet.

It works.

B+ is at a steady 174 Vdc.



EDIT 4:

I'm ditching the power MOSFET idea.

I don't know if it was because I operated the audio chassis at work using the proper voltages or if the MOSFET was just a bad idea as I lost two of them. One apparently shorted and one apparently wouldn't pass any voltage.


So it looks like I will go with the 2N3902 and the other power transistor in the two transistor regulator circuit I posted, although I will need to now order a 180 volt zener since the other one shorted.

That said is it possible to use the voltage reference tubes as the 180 volt regulator?


Top
 Profile  
 
 Post subject: Re: Collins R-390 solid state regulator
PostPosted: Oct Mon 30, 2017 4:13 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 17981
Location: Warner Robins, GA
I tried the circuit using the 2N3902 as the pass transistor and the BD-159 as the driver.

All tests were with a B+ supply of 250Vdc.

Using the voltage reference tubes with a 10K resistor in series I get 180 Vdc unloaded.

Loaded with nearly 200mA the output is 174Vdc.

If I remove the 10K the voltage is 167Vdc unloaded and drops less than a volt with a nearly 200mA load.

So a 180 volt zener is what I need and the circuit should then work properly.

I don't like that there's no feedback, but so long as the circuit meets spec which is +/- 1Vdc I'll be happy.

Here's the current schematic.

Attachment:
Regulator 5.jpg
Regulator 5.jpg [ 64.15 KiB | Viewed 335 times ]


Provided I don't have a 180 volt zener and have to order one I will also order the proper 1K resistor so that I can load the regulator to 180mA just to be sure its output voltage stays within +/- 1Vdc.

Will also be checking how hot the heatsink gets with an input B+ of 250Vdc which is subject to change based on the heatsink temperature.

Here's what I get with no load on the supply using the voltage reference tubes in series with a 10K resistor.

B+ 250Vdc
First regulated voltage 200.8Vdc.
Final regulated voltage 178.99 Vdc.
Output voltage 178.629 Vdc.

Here's a couple pictures of the regulator.
Attachment:
20171030_145432.jpg
20171030_145432.jpg [ 190.67 KiB | Viewed 331 times ]


Attachment:
20171030_145419.jpg
20171030_145419.jpg [ 208.18 KiB | Viewed 331 times ]




EDIT:

I tried two 1N754A 6.8 volt zeners in series with the two voltage reference tubes and here's what I get.

180.37 Vdc no load.

179.3 Vdc at 210mA load.

So it looks like the circuit at least with the zeners and reference tubes will work.

The only problem is this.

The voltage reference tubes will not light up connected in the circuit. I have to disconnect them then reconnect them and they light.

Could it be that the slow rise in voltage caused by the 22uF cap is the problem?

If so I may switch the 22uF and 2.2uF caps around and try that.

Keep in mind using the reference tubes is only for testing purposes, but when I order the 180 Vdc zener I will test both and see which has the better regulation from no load to full load and use that.

I know my test setup is a little flawed as I have to increase the AC input to the power transformer so I get 312Vdc unloaded out of the transformer before I connect the load at which point the main B+ drops down to 250Vdc.

Don't know if it is because of the 10uF filter cap or if the fact that I am using a variable AC power supply to vary the transformer voltage has anything to do with it.

At 210 mA B+ current the ripple voltage is 21.5mVac with each rectifier diode going through a 50 ohm resistor before the first filter cap. Without the resistors the ripple voltage is 23.1 Vac.

My question is will two 47 ohm resistors handle the 200mA B+ current. The power supply has four one in series with each diode and when I swapped to solid state rectification I kept the same resistors connected the same way and used four diodes like originally done.

They are 5 watt resistors and I need four in series between the rectifiers and first filter cap to equal 250Vdc under load.


EDIT 2:

I wonder if I could get any better regulation if I installed another zener before the 200 volt one perhaps rated at 220 volts and use that to feed the 200 volt zener?

I may go ahead and order one just to try it.

The issue is I don't like the full 396Vdc B+ the circuit sees (115 Vac primary voltage) when the tubes are cold being at the B+ end of the 39K resistor as it causes the 200 volt zener voltage to jump up to 212 Vdc. If I did add a 220 volt zener I am not sure yet what series resistor is needed.

Now the circuit itself I believe was designed to work with the original filter cap and also use a 2.2uF cap after the 500 ohm power resistor which would improve the filtering.

I will be ordering two 500 ohm 50 watt resistors so I can series them and have a proper load which can handle the wattage without burning up.

Now doing some calculations using the 500 ohm resistor the article specified between B+ and the regulator circuit I get a voltage drop of 100 volts and the resistor dissipates 20 watts. That will give me a final B+ of 296Vdc feeding the regulator. Based on what I've seen in practice I can go as low as 250Vdc and still get reliable operation.

120 Vac line voltage 296 Vac 416 Vdc B+ unloaded 345 Vdc 210mA load.
110 Vac line voltage 274 Vac 382 Vdc B+ unloaded 320 Vdc 210mA load.

So the 500 ohm resistor would be about right and the other 220 volt zener will not be needed.

Not a fan of a large power resistor dropping the voltage though.

Now if I were to use a 100 volt zener to do the voltage drop what wattage would it have to be?

Mouser only lists a 100 volt zener at up to 5 watts, but I can use 5 20 volt 5 watt zeners to do the voltage drop. That way the voltage drop will be constant and at 5% tolerance with 5 zeners I can have anywhere from a 95-105 Vdc drop which won't be much at all compared to the HV I am using. Doing calculations each 20 volt zener will be dissipating 4 watts at the full rated B+ current of 200mA. I am ordering 8 of the 20 volt zeners as That will give me the ability to try dropping the B+ voltage further to lower the pass transistor's temperature more. The circuit was designed to work with the original rectifier tubes and due to using the diodes my unregulated B+ will be higher.

Attachment:
Regulator 6.jpg
Regulator 6.jpg [ 71.04 KiB | Viewed 318 times ]



EDIT:

Parts were ordered yesterday.

Also realized that I goofed in that I should have ordered another 10uF 450V cap for testing purposes. That's ok though as I can just remove the 10uF first filter cap in the R-390 and use it which is probably better for simulating actual operating conditions anyways.

My only concern is how hot the heatsink will get and what its effect will be on the circuit given the zeners and caps are mounted where they are.

Now in the future if the mod proves to be reliable I may look into replacing the electrolytic caps with film caps which would then have to be mounted external to the heatsink, but there should be room.

I also found out why the reference tubes wouldn't reliably turn on by themselves. each tube has a normal starting voltage of 107 Vdc and a max of 115 Vdc which in series is 214 Vdc - 230 Vdc and given the first regulator is 200 Vdc the tubes would never reach the normal starting voltage.

What I might do is try the 180 volt zener then I'll remove both zeners and try the reference tubes and see which one causes better regulation.

The only problem then is the regulator won't be plug and play which is fine for mine as the heatsink used won't allow an octal tube base to be plugged in, but others may not like the idea of having to mod theirs to make the original reference tubes work.

Now it would be nice if feedback could be easily added to the simple solid state regulator circuit.


Top
 Profile  
 
Post New Topic Post Reply  [ 59 posts ]  Go to page Previous  1, 2, 3  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: ggargia and 7 guests



Search for:
Jump to:  


















Privacy Policy :: Powered by phpBB