| View previous topic :: View next topic |
| Author |
Message |
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Mon 02, 2009 2:31 pm |
 |
|
| Cdoose wrote: | | Yes but the first transformer has to supply the extra power to the other two that are saturated. If they weren't saturated the first transformer won't have to supply so much power. |
Ahhh so...
I see.. gotcha... 
thanks Chuck.
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
 |
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Mon 02, 2009 3:09 pm |
|
|
Ok... we'll get this one way or t'other.
I ordered the dual wound 14v/115 (28 ct) PL 30-28-130
http://www.tamuracorp.com/clie ... X-130B.pdf
I also got a dual 12v/115 (24 ct)
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
radiorich
Member
Joined: 01 Jan 1970
Posts: 5604
Location: tacoma,wa,usa
|
| Posted: Nov Mon 02, 2009 6:44 pm |
|
|
Hello Peter,
I want to keep up on the progress so I maybe build half this amp for my drive-in speaker .
Rich
_________________
 |
|
|
|
|
Tom Bavis
Member
Joined: 01 Jan 1970
Posts: 2587
Location: Rochester NY USA
|
| Posted: Nov Tue 03, 2009 1:01 am |
|
|
| With 90 mA at 190V, there is a total of 17W of DC output. The transformer has to supply this PLUS the filter capacitor ripple current, probably about 25VA, so the 12VA transformer was definitely undersized. 30 VA should be enough, if the voltage works out. |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Tue 03, 2009 3:16 am |
|
|
| radiorich wrote: | Hello Peter,
I want to keep up on the progress so I maybe build half this amp for my drive-in speaker .
Rich |
Hi there Rich:
You'd enjoy it I'm sure.. and it sound terrific too.
The op275 op-amps are cheap and the whole thing makes for an neat little project.
I encourage you Rich!
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Tue 03, 2009 3:27 am |
|
|
| Whammo wrote: | Looks like good work as usual Peter.
My only question is if the tubes and bases were challenging/expensive to procure, wouldn't it have been better to substitute a different tube? |
Yes Whammo... you have a good point.
The fellow who designed this, Karl Redmer from RVR labs, suggested a 6AK6 as an alternative.
I ordered tubes and sockets for them as well and in my early bread-board design on the bench I actually built BOTH circuits and switched between them alternatively for comparrison.
The 6ak6 is a great tube too with a lot less filament current. But since I WAS able to get the 5902s I decided they were so nice and cute too.. so I stuck with them.
But if you wanted to try it with 6ak6s ... that would work fine too!
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
radiorich
Member
Joined: 01 Jan 1970
Posts: 5604
Location: tacoma,wa,usa
|
| Posted: Nov Tue 03, 2009 6:29 am |
|
|
Hello Peter,
Yes it does sound like fun little project .
I really enjoy reading your post on your amp .
Sincerely Rich
_________________
|
|
|
|
|
Cdoose
Member
Joined: 01 Jan 1970
Posts: 765
Location: Hinsdale, IL, USA
|
| Posted: Nov Tue 03, 2009 6:31 am |
|
|
Peter,
Do you have another Triad 25 V transformer like the one you're using?
I just did a couple quick calculations and it looks like if you were to use another Triad instead of the Tamura, you'd be able to get about 130V +B.
The audio power will go down a bit but not much. the power required for the four 5902s will be about 15 Watts (the plate current will be about 18 mA per tube) and with a couple more watts for the op amps/Zener it will be about 17 Watts total DC power required.
You won't have to change the cathode bias resistor you can leave the 420 ohm in there.
So as I see it you'll have 11 watts for the heaters, and about 20 watts for the 5902s and op amp. The Triad transfomers are rated at about 50 VA, figure 90% efficiency and they will be loafing along at 35 watts. I'd give this a shot and see how it works. This way you won't be saturating either transformer.
_________________
Chuck D. KB9UMF |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Tue 03, 2009 9:39 am |
|
|
| Cdoose wrote: | Peter,
Do you have another Triad 25 V transformer like the one you're using?
I just did a couple quick calculations and it looks like if you were to use another Triad instead of the Tamura, you'd be able to get about 130V +B.
The audio power will go down a bit but not much. the power required for the four 5902s will be about 15 Watts (the plate current will be about 18 mA per tube) and with a couple more watts for the op amps/Zener it will be about 17 Watts total DC power required.
You won't have to change the cathode bias resistor you can leave the 420 ohm in there.
So as I see it you'll have 11 watts for the heaters, and about 20 watts for the 5902s and op amp. The Triad transfomers are rated at about 50 VA, figure 90% efficiency and they will be loafing along at 35 watts. I'd give this a shot and see how it works. This way you won't be saturating either transformer. |
Hi Chuck!
...Great minds think alike!.
That's what I did already Chuck-Sir...lol
At first I took the two Tamura 16v transformers and wired them in an interesting combo of series and parallel such that it was 8-8-8 = 24v
The middle two 8v windings are in parallel and the ends are series.
Anyway.. it worked fine too!... lol
And sure enough the load on the Triad dropped right down to a nice cool comfortable level.
But it only provided about 137v for B+
So I abandoned that scheme and that's when I put another Triad in there. I wanted to see how the new Tamura dual wound14v (transformers that I ordered today would perform voltage-wise.
I was worried that I might have a hard time getting 150vDC or so out of a 28 volt transformer being fed 28 volts.. .. figured I'd only get 115vac out to the bridge.... right?
That's why I also ordered a 12/12v (24v) variety too.
Of course I won't be keeping the big 2 amp Triad in there because it's physically too big for the space I have.
The new 28v Tamura looks small enough to fit though.
With the Triad as it is now.. I am getting 29vAC input and 117vAC output.
So it's a struggle to get up to a solid 150vDC for B+
With a 330uf input filter cap however I am just able to get 152vdc at the bridge output with 1-1/2v to 2v PP ripple.
So I replaced the next 220 ohm resistor with a 50 ohm feeding the next 330uf cap.
This 50 ohmer drops only 4 volts now leaving me with 148vDC @ 150mv PP ripple.
This is just marginally acceptable.
I tried paralleling an additional 300uf cap as a test on either side of the pi-filter but the voltage stayed the same. I could only get the ripple down to 100mv PP.
So this will be workable, yes... but a bit higher would be better no?
So I'll try out both the new 14v ( 28 ) and the 12v (24) transformers to see what I can get.
Perhaps the 24v one won't be stretching the parameters too much.. and then I'm sure I'll have enough voltage.
What do you think Chuck?
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
Cdoose
Member
Joined: 01 Jan 1970
Posts: 765
Location: Hinsdale, IL, USA
|
| Posted: Nov Tue 03, 2009 1:36 pm |
|
|
Hi Peter,
Glad to hear you tried the different transformer combinations.
Pesonally if it were me, I'd just go with the lower +B of about 130V. You'll have less power to dissipate in heat and not much difference in audio output power.
With 155V +B the four tubes require about 22 watts at maximum audio output of about 5 watts/channel.
With 130V +B the power required is 15 watts at 3 watts/channel.
That's only about a -2 dB reduction in audio output power, I doubt you'll notice it much.
This is where an adjustable benchtop DC powersupply comes in handy to try different +B at the turn of a knob.
Give it a shot with the new Tamura transformers. I think the 24 V ct transformer will still run a bit hot. unless they put in more steel than is required for normal operation....unlikely.
BTW, my calculations show the ideal plate-plate OPT impedance to be 5k, but if I remember right you've found 8k to be better?
Have you tried 5k for the OPT impedance how does it sound?
This is fun stuff isn't it?
_________________
Chuck D. KB9UMF |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Wed 04, 2009 4:02 am |
|
|
Hi Chuck..
I'd like to understand the math you did there about wattage out. I don't know how to do that.
Is it possible to explain it a bit?
But I do see your point. .. hardly matters.
BTW... as I measure B+ today after an hour running.. it's right at 150v.
Yesterday was 148v.
As far as the 5k taps.. I can't be sure now what I thought was best.
I tried every tap and some were VERY obviously bad but others from 3k to 8k were all too similar.
So I picked taps 2 & 5 that Karl used in his print (8k for an 8-ohm speaker load)... and it seems fine.
But it's hard to move the taps now as the transformers are too close together for me to easily switch them.
So I'll have to run another test with a transformer removed.
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Wed 04, 2009 8:44 am |
|
|
Well Chuck... I guess I must be getting dead around the ears....lol
I switched taps back and forth between (1-5 or 2-5) 5k and 8k and to be honest I couldn't tell the difference.. I don't think.... lol
I mean ...sometimes I thought it sounded "better" sometimes... one way then louder the other way .. and sometimes sort of "richer"? ...but I can't tell.
I even tried 3-5 taps which is 17.5k ohms ... and that was pretty darn good too... lol
So how the heck can I get a real definitive test to determine optimum settings for load impedance?
Is there a way to measure power/efficiency or whatever?
Will the scope see a difference on the plate voltage in the PP amplitude with different taps? ( using a 400hz test tone)
I'm using a couple of ole 1970's stereo speakers.
8 ohms.
Iinside there's an 8" woofer in parallel with a 2.2uf cap to a 2" tweeter.
This is supposed to "crossover" at around 7,000 hz. I believe.
(does that keep the impedance at 8 ohms like that?)
Does the cap as a crossover in there change the circuit to look like two 8 ohms in parallel = 4 ohms? or does it remain 8 ohms?
Update answer:
"The crossover allows you to place two 8 ohm speakers in parallel, one for the highs and one for the lows, and still present an 8 ohm, not a 4 ohm load to the amplifier (which is what the amp would see without the crossover network)"
Guess that answers that!
Anyway... The sound seems "darn good" too.... which-ever way it's tapped... lol
BTW ...("Darn good" is one of them there scientific terms.)
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
Cdoose
Member
Joined: 01 Jan 1970
Posts: 765
Location: Hinsdale, IL, USA
|
| Posted: Nov Thu 05, 2009 5:46 am |
|
|
| Quote: | Well Chuck... I guess I must be getting dead around the ears....lol
I switched taps back and forth between (1-5 or 2-5) 5k and 8k and to be honest I couldn't tell the difference.. I don't think.... lol
I mean ...sometimes I thought it sounded "better" sometimes... one way then louder the other way .. and sometimes sort of "richer"? ...but I can't tell.
I even tried 3-5 taps which is 17.5k ohms ... and that was pretty darn good too... lol |
Peter,
I know what you mean. I was trying to hear the difference in some 12AX7 preamplifier circuits, and I really had to make the distortion bad to hear a difference. The mid range is pretty well shot on my ears.
One way I check for a good OPT match is the following:
Hook up an 8 ohm load instead of a speaker. Connect a signal generator on the input to the amp. I think yours put out 400Hz?
1.) Look at the control grid signal with your scope and bring the generator signal up until you see clipping at the grid. Back off the signal level until you don't see any clipping. Measure and record this signal level with your scope or a good AC meter. For the next tests don't exceed this input signal.
2.)Now monitor the signal across the 8 ohm load, bring the generator signal level up to the level from step 1. If the OPT impedance is matched properly, you should see a relatively clean undistorted sine across the 8 ohm load with the control grid signal level just shy of clipping. If it's distorted try different taps on the OPT to get the best looking sine. Make sure you don't operate the amp at high power without a load hooked up.
Another way you can measure distortion without an distortion analyzer is to use the scope in the x/y mode. The input signal goes to the x channel and the output signal to the y channel. Set the scope sensitivities to get a diagonal line with the signal present. When the output starts to distort the line will become curved.
You can calculate the output power by taking the square of the RMS voltage across the load divided by the load resistance . i.e. Power = V^2 / R So if you measure 4 Vrms across the 8 ohm load, you'll get 4*4/8 = 2 watts
You always want to do distortion and power measurements with a purely resistive load, not a real speaker because of the uncertainty of the actual impedance at the measurement frequency. You know the impedance of a speaker varies like crazy depending on frequency.
Not sure on the impedance of the two speakers with the 2.2 uf cap. I'll have to think on that. My gut tells me that the impedance will change a lot more than a single speaker would. Seems to me the net impedance would be less than 8 ohms for most frequencies, but I have no way of knowing the characteristics of the two speakers.
Getting back to your question on the math involved for the push pull output. It's really beyond what I can put down in a few paragraphs. Traditionally you use the tube spec sheets to use the values already figured out by the manufacturer. That's fine if the spec sheets show the same operating point as you want to use. If not you must use the plate characteristic curves, plot a load line for a particular load, then determine the plate voltage and current swings for the chosen load line. This is relatively easy for a single ended output. It's much more involved for push pull because you have to use two sets of plate curves line them up properly then figure out what they call a composite set of curves which represents how two tubes behave when hooked up in P-P.
Anyway this is all very tedious and confusing the first few times you do it, and if you haven't done it for a while you'll forget how.
So since I'm basically lazy I created an Excel spreadsheet to do all the calculations and plots pretty much automatically. All I need is an image of the plate curves for a particular tube, and then I can figure out how that tube should behave for a SE or PP common cathode circuit. It will work for fixed bias or cathode bias. Again it's way too much to cover right now. I've been working on this stuff off and on for the past two years. I'm planning on writing up an explanation of how this all works and putting it on the web some day, but right now I've got too much stuff going on in my life.
Long story short, if you give me the tube type and +B I can tell you the maximum plate current, max power output, power input required, cathode resistor or fixed bias value, optimum OPT load for a SE or PP output circuit, as well as the plate curves and a slew of other useless info. I've found the calculations agree with the tube spec sheets to within about +-5% and on tubes I've measured myself it's a bit better. The accuracy is pretty much dependent on how well the curves represent the real tube.
Hope I didn't bore you with this long winded post.
_________________
Chuck D. KB9UMF |
|
|
|
|
radiorich
Member
Joined: 01 Jan 1970
Posts: 5604
Location: tacoma,wa,usa
|
| Posted: Nov Thu 05, 2009 7:17 am |
|
|
Hello Peter,
I want to add what Cdoose Said.
Pick your self a couple of 8-Ohm Non-Inductive Resistor
from radioshack they are good for 20 watts to .
http://www.radioshack.com/prod ... Id=2062288
See a regular resister has inductance atleast a wirewould does and that will cause problems with you readings on the scope.
Rich
_________________
|
|
|
|
|
radiorich
Member
Joined: 01 Jan 1970
Posts: 5604
Location: tacoma,wa,usa
|
| Posted: Nov Thu 05, 2009 7:21 am |
|
|
Hello Peter,
one thing if I build one of these amps here at home .
I can come up with most of it .
Without buying to much .
Mainly the tubes and maybe the preamp,
depends if I have any more of those left in my stock.
Rich
_________________
Last edited by radiorich on Nov Sun 08, 2009 12:08 am |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Thu 05, 2009 7:44 am |
|
|
| Cdoose wrote: |
Peter,
One way I check for a good OPT match is the following:
Hook up an 8 ohm load instead of a speaker. Connect a signal generator on the input to the amp. I think yours put out 400Hz?
1.) Look at the control grid signal with your scope and bring the generator signal up until you see clipping at the grid. Back off the signal level until you don't see any clipping. Measure and record this signal level with your scope or a good AC meter. For the next tests don't exceed this input signal.
2.)Now monitor the signal across the 8 ohm load, bring the generator signal level up to the level from step 1. If the OPT impedance is matched properly, you should see a relatively clean undistorted sine across the 8 ohm load with the control grid signal level just shy of clipping. If it's distorted try different taps on the OPT to get the best looking sine. Make sure you don't operate the amp at high power without a load hooked up.
Another way you can measure distortion without an distortion analyzer is to use the scope in the x/y mode. The input signal goes to the x channel and the output signal to the y channel. Set the scope sensitivities to get a diagonal line with the signal present. When the output starts to distort the line will become curved.
You can calculate the output power by taking the square of the RMS voltage across the load divided by the load resistance . i.e. Power = V^2 / R So if you measure 4 Vrms across the 8 ohm load, you'll get 4*4/8 = 2 watts
You always want to do distortion and power measurements with a purely resistive load, not a real speaker because of the uncertainty of the actual impedance at the measurement frequency. You know the impedance of a speaker varies like crazy depending on frequency.
Not sure on the impedance of the two speakers with the 2.2 uf cap. I'll have to think on that. My gut tells me that the impedance will change a lot more than a single speaker would. Seems to me the net impedance would be less than 8 ohms for most frequencies, but I have no way of knowing the characteristics of the two speakers.
Getting back to your question on the math involved for the push pull output. It's really beyond what I can put down in a few paragraphs. Traditionally you use the tube spec sheets to use the values already figured out by the manufacturer. That's fine if the spec sheets show the same operating point as you want to use. If not you must use the plate characteristic curves, plot a load line for a particular load, then determine the plate voltage and current swings for the chosen load line. This is relatively easy for a single ended output. It's much more involved for push pull because you have to use two sets of plate curves line them up properly then figure out what they call a composite set of curves which represents how two tubes behave when hooked up in P-P.
Anyway this is all very tedious and confusing the first few times you do it, and if you haven't done it for a while you'll forget how.
Hope I didn't bore you with this long winded post. |
Thank youi very much Chuck:
No you didn't bore me at all.. in fact I thirst for more... sort-of.
I'd like to actually do that plotting with someone like you looking over my shoulder sometime just to actually "experience" the correct (if tedious) method... such that I might be able to rely on myself to do it alone when desired.. y'know?
But just like you built your own spreadsheet for this ... you'd think that in this age of audiophile wildness and on-line calculators for everything... that there must somewhere be thw ideal EASY tool to use on line for this.
After all look at all the stuff that a program like "bass box-pro" software figures out by comparison... no?
If they can do that there must be something for PP parameter calculations.
How do I know if I'm operating PP now in A, AB1 or what?
Nothing is ever easy...lol
As far as feeding in my 400hz signal generator signal... then looking for distortion.... lol
My signal generator does produce a reasonably nice 400hz sine wave!.... until it's loaded feeding the amp.
Then it looks all weird.
It looks like some weird-looking waveform. So how can I see distortion? I have to compare weird to weird and hope I can see the exact areas if distorted.
Let me go take a few shots to show you.
BRB.
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Thu 05, 2009 8:01 am |
|
|
Ok. here's the signal before being fed in.
Then it has to be cranked up or the input loads it down.
But then the signal is all weird-shaped.
And then the inverted and nonp-inverted op-amp out signals.
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Thu 05, 2009 8:03 am |
|
|
| radiorich wrote: | Hello Peter,
one thing if I build one of these amps here at home .
I can come up with out buying to many of them to.
Mainly the tubes and maybe the preamp,
depends if I have any more of those left in my stock.
Rich |
HI rich:
If you have a PC-board layout program maybe this would be a nice easy project for it.
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
radiorich
Member
Joined: 01 Jan 1970
Posts: 5604
Location: tacoma,wa,usa
|
| Posted: Nov Thu 05, 2009 8:42 am |
|
|
Hello Peter,
Well Those wave forms look good
you sure do nice photo taking  .
Far as having a PC board layout program I have cad lite .
It will give me a reason to buy one of those programs.
Rich
_________________
|
|
|
|
|
Pbpix
Member
Joined: 01 Jan 1970
Posts: 3442
Location: Haledon, NJ, usa
|
| Posted: Nov Thu 05, 2009 9:09 am |
|
|
| radiorich wrote: | Hello Peter,
Well Those wave forms look good
you sure do nice photo taking .
Rich |
Thanks but not really so good.
Just real quick hand-held shots.
....so a little blurry from camera shake as it was a difficult angle to hold my arms out at..
The trick is just to turn off the flash.
_________________
Peter G. Balazsy
WB2eqt
www.pbpix.com/radio
"Whoso would be a MAN... must be a non-conformist ...for nothing is at last sacred as the integrity of your own mind."
... (Emerson) |
|
|
|
|
|
