| andrew_whitham |
Hi,
I think you might have made an error in your schematic. Its a push pull amp right? The 6n1p has to be connected to two el34s maybe at pins 6 & 8? (I guess its one 6N1P per channel?)
Regarding tube substitution; the Russian types are supposed to be rather good to start with, so the benefits might be a bit slight, BUT
If you havent already, have a look at the Duncan amps datasheet locator,
http://tdsl.duncanamps.com/tubesearch.php
Although substitutons seem to be a bit thin the pinout is quite similar to the ECC8* types and they might provide an alternative there, depends what you want to achieve.
Good luck!
Andy |
|
|
| dsavitsk |
Is it me, or is this all weird. Sophia Electric owns a trademark in the phrase "S.E.T." (the serial number is 78163221 which can be looked up at the USPTO, or just search for Sophia Electric) for use with audio equipment which means a) only they can sell audio equipment labeled "S.E.T." and b) it does not mean Single Ended Triode. Instead, it means nothing. So, this amp, called "S.E.T.™ Music EL-34" is actually a push pull amp and is so called because "S.E.T." does not mean anything in this context.
Sorry to thread jack. This just struck me as odd. |
|
|
| Svein_B |
Yes the S.E.T. trademark is a bit weird.
In any case I do not really think the cheap China amp is a copy of the Sophia Electric, it is probably rather the Sophia which is a re-branding/OEM of a Chinese amp.
The circuit drawn is obviously not yet correct, but tracing schematics from the live amp is a good exercise.
The 100nf cap is probably connected to the other end of the 300K resistor.
The phase splitter might be a basic paraphase, with the grid of the second triode half being driven by the 330K/20K divider.
The anode of the second triode half (pin 6) will probably go to the other EL34 grid via a 100nF cap.
And I guess there will be some resistor from pin 8 to gnd.
SveinB |
|
|
| Cobra2 |
output
the resistance reading for the trafo seems odd?
Arne K |
|
|
| Svein_B |
| quote: | Originally posted by Cobra2
output
the resistance reading for the trafo seems odd?
Arne K |
Well, maybe not so bad. These are DC resistance readings.
I would expect the secondary to be a little higher, probably more like .5 ohm. Since you have it apart, you should measure also the AC ratio, connect an AC signal to one side and measure on the other. The square of this will be the impedance ratio.
Your diagram is looking better now. The 100nf still need to be connected to the other end of the bottom 330K. Please also check the feedback connection, it is a chance it shouild be connected to the 900/90 ohm cathode resistors.
SveinB |
|
|
| Cobra2 |
Latest corrections...I cant do better...?
B+ unloaded ~ 450V
Arne K |
|
|
| Cobra2 |
attachments dissapears...?
Arne K
@ SveinB, you are right about the feedback (not corrected yet on drawing) |
|
|
| mississippi |
| When you make some clear photo's of the upper and inner side of your amplifier and post them here or mail them to me, then I will make you a nice drawing of that amplifier. |
|
|
| Cobra2 |
2
(@ mississippi, if you mail me, I can return better pictures)
Arne K |
|
|
| mississippi |
| you have got mail! |
|
|
| tkifowit |
The schematic is hard to read.
As far as what you can do to make it sound better,
first; throw the chinese tubes away (i like holland tubes - tube hop)
second; a real good output transfomer (i like hammond)
last; replace all in signal caps with good ones.
Most Chinese knock offs are based on very good designs just cheap parts.
It might surprise you. I do a bunch of these for freinds. |
|
|
| Cobra2 |
PSU is a bridge (4x1N4007)and a 150uF/450V cap, splits to each channel, thru a 50 ohm resistor (to each ch) and another 150/450 cap.
The feedback point(resistor) should be connected to the point between R9 & R11.
The 330k resistor between the tubes probably go to gnd, and not between the 10k...
I'm new to tube-audio topologies, and have trouble finding anything similar, all I have seen, have trafos with more taps...
Arne K
@ mississippi, please mail me again, I sent & deleted...and mail bounced (mailbox to small) |
|
|
| mississippi |
| You have got mail :D |
|
|
| Svein_B |
| quote: | Originally posted by Cobra2
The 330k resistor between the tubes probably go to gnd, and not between the 10k...
|
The 330K resistor may be drawn correctly. I think it does not go directly to GND, but via the 20K/10K (trimmer pot?) in order to drive the grid of the second triode half. This is the very old-fasioned paraphase phase splitter used by Yarland and other cheap Chinese amps. It work OK, but needs adjustment when "rolling" driver tubes to get the correct AC signal balance.
Modifying the driver/phase-splitter may be an idea, but I suspect that gain might be marginal with only one double triode per side to drive EL34, so options may be limited.
SveinB |
|
|
| Cobra2 |
There are just one trimpot...
Arne K |
|
|
| mississippi |
| Here is the schematic diagram of this amplifier. |
|
|
| SY |
Whenever you see output tubes on a cheap circuit board, and especially a board that's not highly perforated, you know you've got an amp destined for progeria.
As for the overall design, cough, cough, it's not exactly what I'd do... |
|
|
| Cobra2 |
Can fix holes, both in pcb & chassi...
Next tip please!
Arne K
1000 x thanks to mississippi for making the great drawing...
|
|
|
| mississippi |
Youre welcom.
and for| quote: | | As for the overall design, cough, cough, it's not exactly what I'd do... |
I'll agree! |
|
|
| tkifowit |
I always like seperate transformers/rectifiers for low and high voltage.
You can keep the transformer you have now for high and just add another for the heater. Not real exspensive. |
|
|
| SY |
| I guess you have to first define what you're after. A test bed for learning and experimentation? A long-term amp to use unmolested for listening? Different answers to those questions. |
|
|
| Cobra2 |
Since it barely can make a pigs ear, I see no hope as a silk purse...
It will mostly be a test-bed,(training/learning), but first; it has to be usable.
As it it(was); it has no bass or punch...and had a glary top-end.
And not even close to a varm midrange.
Arne K |
|
|
| SY |
| I'd spend some time measuring the output transformers. If they've got the bandwidth and power handling to make a decent amp, then it could be worth stripping everything else out and starting with that as a base. If they don't, just reconnect them, sell the amp, and use the money to buy some decent iron. |
|
|
| Cobra2 |
Bandwith I can check...how do I check power-handeling?
Arne K |
|
|
| SY |
| Run bandwidth at different powers while observing waveforms for distortion. You're most likely to see problems at the low end if the iron is light, high frequency problems if the winding is poor, both if you're dealing with a really bad specimen. |
|
|
| Cobra2 |
My "lab" is not exactly set up for high-power testing, I could only get 10V out of my generator. (~250mV over load)
With a 8 ohm load, a sinus looke fine from 10+Hz to 23kHz, started dropping, increased to a resonance-peak at 25kHz, and fell thereafter quite fast...
With square, it looked good from 10-5kHz, then it showed a small ringing, squares got curvey upward...still "square" past 15kHz then got fairly out of any shape...
Does this tell anything about the trafo? (other than sh**?).
Arne K |
|
|
| SY |
Don't be afraid to use that output stage for power testing.
If you really only get 25kHz and peaking from the transformer at small signals, that's an indication that trying to modify this amp is like polishing a turd. |
|
|
| Cobra2 |
I really guessed the trafos were the cheapest part...
There are space for replacement...
So, what should I look for / where? Price-range? (no, not the gold-plated, silver-wired-ones).
(offers welcome)
Arne K |
|
|
| SY |
| James, Dynaclone, Magnequest, Sowter, and Lundahl immediately come to mind. |
|
|
| Cobra2 |
I know the brand-names...and a few more, but, what properties should I look for? Ohm/ratio/power/?...
BTW, are the stock trafos usable for anything? heater-supply?:clown:
Arne K |
|
|
| SY |
Start from the beginning. What performance parameters are you looking for? Power, especially.
The power transformer might be OK; it would be helpful to measure the output voltages no-load and under load, then run it loaded for a few hours to measure thermal rise. A resistor load is good enough for this sort of test. |
|
|
| Cobra2 |
It is rated 2 x 35W (in China-watt) no idea about distortion-level.
30-40W should be reasonable to get?
I ran it hard/LOUD for an hour, into speakers(8 ohm/90dB), powertrafo did not get hot...maybe 5 degree rise from idle, or less.
So I guess it can stay...
Arne K
looking at specks, typical pp 2 x EL34 trafos have Raa from 750-5k, lower for higher power? How does this influence other parameters? Need higher Volt/Amp? |
|
|
| SY |
I'd look for something 4k3-6k6 plate to plate, preferably with ultralinear taps (whether or not you end up going that route, it's best to be flexible). Appropriate power rating. In that power and impedance range, you'll want to consider EL34, 6L6-oids, and maybe paralleled EL84. There are other choices, but these are tried, true, and come with decades of documentation.
The choice of 9k p-p for a pentode EL34 amp is not something I understand. But this designer is working under a set of goals and constraints that I don't and can't know, so maybe there's some logic to it. Maybe not.;) |
|
|
| Cobra2 |
The most common logic here is $ ...
Thanx a lot...now I have something useful to work from.
Maybe it will transform into something slightly nicer than a pigs ear after all... :D
Arne K |
|
|
| Cobra2 |
here...(a couple of resistors has been tweaked slightly)
Arne K |
|
|
| Cobra2 |
It sounds much better, and interestingly, the Electron-tubes(chinese) has more bottom-end-"omph" than JJ EL34
Midrange & top-end is much improved.
The stock 6P1 and Sovtek 6P1P did less difference.
Next...trafo-shopping... - should I go "UltraLinear"?
Arne K |
|
|
| Svein_B |
| quote: | Originally posted by Cobra2
here...(a couple of resistors has been tweaked slightly)
Arne K |
Well, this looks like more than a little tweaking of resistors :bigeyes:
You seem to have picked up on tube amp design skills really fast and re-done the whole driver/phase-splitter topology.
A couple of questions out of my own curiosity:
The connection of the concertina phase splitter's cathode resistor seem a little unorthodox, could you explain?
Why did you add the input cap?
SveinB |
|
|
| Cobra2 |
The credit for circuit change and drawing should go to mississippi...
(but I skipped that cap ;-) - and gave the input-stage slightly more bias...compared to last drawing.
Phase-splitter is very similar to last figure here:
http://www.arrakis.es/~igapop/russi...to%20use%20them
Arne K |
|
|
| andrew_whitham |
Cobra2 - from our discussions on the other, more tranformer based thread.
I got curious and plugged the Numbers here into the TCJ p-p calculator (so all this is for operation as Triodes IMO the most musical use for an EL34 anyway :))
9k2 opt? Oh well here goes....
"Should" (as a triode) get you 11w output but leave class A at 4.24 watts
reducing the cathode bias resistors on the output stage to say 350R (in total) makes it Class A all the way, but power drops to 9.88w
Not so bad as a loss. Also check the bypass caps, I have 1000u in mine, needed 600 odd on paper so 330U is a bit light? (the parts bin could only supply 1000u so thats what I used.)
incidentally your higher than expected trafo impedance is the second time I've seen that done. Ok for a Triode its lower distortion at the expense of power but not for a pentode... where it should, er, suck. (more distortion & less power) Hmmm odd
Actually looking at your circuit, have you tried triode mode? Its easy to do. Add a screen resistor though, maybe a couple hundred ohms
Lastly try messing with the feedback loop, Ok I have an SE amp but it REALLY worked once I got it somewhere close. Only a question of clipping in resistors. BTW adding a cap in parallel with the feedback resistor didn't help in my case. I tried a range of values hoping for a subjective bass boost - nope!
Andy |
|
|
| Cobra2 |
No wonder it lacked power...
I did not try triode-mode, (was looking in to it, but I needed all the power it had, and more, for my use/speakers).
I was however tempted to see if I could use 2 more EL34(in paralell) and in triode-mode, to get some more power.
But then again, the power-trafo will need an upgrade.
Maybe cheaper than 2 new op trafos...
Will this idea work with the op trafos I have?
Arne K |
|
|
| andrew_whitham |
OH hang on
I have a question! whats going on with the feedback loop fed into the gain tube's cathode. Ok with it so far, but what appears to be a concerina phase splitter is sat on top of that. is this a different phase splitter or shoud the 33k resistor teminate at ground not the hot end of the 180ohms?
(and does it matter anyway?)
If anyone can clear that up I'd appreciate it - It just looks asymmetrical to me, probably nothing and blindingly obvious but still...
Andy |
|
|
| andrew_whitham |
| quote: | Originally posted by Cobra2
No wonder it lacked power...
I did not try triode-mode, (was looking in to it, but I needed all the power it had, and more, for my use/speakers).
I was however tempted to see if I could use 2 more EL34(in paralell) and in triode-mode, to get some more power.
But then again, the power-trafo will need an upgrade.
Maybe cheaper than 2 new op trafos...
Will this idea work with the op trafos I have?
Arne K |
you get waaay more power in Pentode mode as shown than with triodes.
You might be surprised how little power you need. After a diet of 50w transistor amps with the attenuator at 9 oclock I was surpised to learn that didn't mean 25% power (log pot!) I measured this, and found that at my most comfortable listening level I was using only 3-5 watts even if the amp could go MUCH louder
I reckon on about 3w from my SE amp - it goes loud enough. OK not as loud as the prodigy in my lounge but loud enough - 10w of triode ought to be do-able. I have 89dB Castle Harlech's But I don't tend to listen at ear crushing volume either
going to parallel tubes is a nightmare! All trafos need to be comfortable with 2x the current and the requirement for the impedance is halved - your 9.2k OPTS are really in the wrong direction then.
Not saying dont though" Until recenly I was planning on an Octet of KT88's then one day I sat myself down and said, "now come on be serious" :)
Andy |
|
|
| Cobra2 |
Well, by trial, I need power (WATT)... My speakers are not that efficient, and my livingroom is 60-70 m2.
And I already anticipated your answer about doubling/paralell...
Arne K |
|
|
| andrew_whitham |
| quote: | Originally posted by Cobra2
Well, by trial, I need power (WATT)... My speakers are not that efficient, and my livingroom is 60-70 m2. |
Ok I see you need Juice! All I can suggest then is UL (which I *think* you can do with a MOSFET if your OPT's dont have the tapping? anyone? I have seen this I'm sure, but I cant remember where!
course its open to debate, but yes parallel push pull would seem to be the way to go (anticipated or not ;-)) Parallel PP KT88's should get you three figure numbers! Outside my expertise to
even start to comment on though.
Andy |
|
|
|
