looking for very good SE output transformer

Not only kits, but apparently "complete" products coming from China require minor or major rework. You can find what I did for two tube amps here:

http://www.diyaudio.com/forums/tubes-valves/208987-yaqin-mc-100b-powerful-advertised-value.html
http://www.diyaudio.com/forums/tubes-valves/255409-lian-845-set-kit-commercial-product.html
I read your first two posts in the first thread. That does not sound bad at all already. The kit I built is not useable in the original form. I had to put in a cathode follower to make it usable.

You talked about the SRPP driving 1Mohm. I can see the reason if it is driving the power tube. this is the same reason why the cathode follower I added actually make the kit amp that is ready to go to the garbage can to at least usable. When the grid of the power tube approach 0V or going +ve, it draws grid current and totally distorts the input waveform of the driver if it doesn't have a low impedance drive. I reduced the THD by half adding the cathode follower. Cathode follower is just the primitive way of SRPP, idea of both are to provide low impedance drive. Of cause SRPP gives higher gain.

These kind of problem and over rated power is common in SS amp even by big manufacturers. I studied a lot of SS schematics, even amps from Yamaha, Onkyo and the like push also. SS amp might be able to give out rated power.....BUT how long can it last? You see 100W amp with one pair of output transistor mounting on small internal heatsink!!!! I wonder how long it takes to burn the amp if you really push it.

Even if it does not burn, THD will get really bad because the output transistors go into beta droop long before.

Thanks
 
Last edited:
Actually second article is more relevant to your case. Makers stated 25W, and deliver less than 5 without mods. Rework allowed getting 13W with reasonably low distortion.
 
Excuse me of being redundant, I can't find on the datasheet the max power of EL34, I only calculate from their spec to 23W( I take 250V - 20V of cathode voltage). So you confirm the max power dissipation of EL34 is 25W. That I just calculate the power by choosing the current for a given power supply voltage.

The datasheet gives a few operating points for 8-11W output with 10% distortion, but that's pentode-connected. Triode connected is more like 6W with 8% distortion, but that's with a B+ of 375V, and all of the above depends on the appropriate load of course.
 
Yes, use separate source of bias adjustable for required idle current. Generally in SE design, idle current is chosen at max allowed dissipation for power tube. In this case it will be 25W. Triode mode gives you a bit less power, but less distortions too. Ultra-linear circuit is unusable without negative feedback. Also 300 volts is too low for EL34, you need to use 500-550 volts at plate for better results. Of cause transformer should present higher plate load in this case too. If you cannot get higher voltage, use KT90 tube biased at 120ma (it allows 45W dissipation). KT series of tubes works better with low plate voltage (like your 300V). For power supply use solid state rectifier along with choke to get best ripple rejection.
After reading your post, I was thinking about higher +B, I went and pull the rectifier tube, put in the 2 1N5408, I got 342V +B!!!!

I did the sine wave test it looks good. Too bad with 250ohm cathode, the plate current is 83mA. For 342V-20V=322V, the power dissipation is 322V X 0.083A = 26.6W. It's over the limit.

So I change the cathode back to 333ohm and I got 70mA. +B gone up to 360V. So power dissipation is 336V X 0.07A=23.5W. This is safe.

I measure THD on 1KHz and 20KHz at 10Vpp, THD is down more than half from using rectifier tube. it's less than 2% @1KHz, 3%@20KHz. Seems respectable to me.

I am so excited I have to post this first before going onto sound test. I get 12Vpp useful voltage driving the speaker. This is much better. I expect to hear improvement. The test of yesterday make the amp goes from dumpster material to at least listenable. But still quite a ways off from my SS amp. Hopefully this will move it up.


Maybe I should think about 2 tubes in the future, go 400V or over and run total of about 120mA so each tube run 60mA and dissipate 23W.



My only problem now is before the power tubes start to conduct, +B is 460V, that's over the voltage rating of the caps.
 
Last edited:
Sadly, the sound does not improve as much as the the scope and QA401 show. Maybe a slight improvement, but nothing to write home. I think it just need more power and better high frequency response. It still doesn't have the dynamics, the richness and the sound stage as my SS amp. At this point, I think the -3dB at 20KHz is really choking the amp and there's nothing I can do except changing the OPT.

So far, still the cathode follower safe the amp from going into the dumpster.
 
Back to OPT

I look at some of the suggestion, I group some into two groups

1) Cheaper ones below $100: Edcor and Hammond. I know the CXSE line of Edcor is $95, I don't know about Hammond.

2) Over $250 to $350: Sowter, Electra Print, Lundahl.

I want to hear from someone that use both the expensive and the cheap and tell me their impression. Is it really worth the extra money?


Also, if I just change the OPT in my cheap amp, would I expect to hear a difference ( of cause I mean improvement!!!)?


Also, if I buy new OPT, I would like to buy one that can use two tubes in parallel. The primary of the cheap transformer should be about 4K from calculation. If I buy one for two in parallel, I should buy one with 2K output impedance ( 2.5K is the closest I see). Would I degrade the sound if I put the new transformer into the cheap amp because it is miss matched?
 
Last edited:
Back to OPT

I look at some of the suggestion, I group some into two groups

1) Cheaper ones below $100: Edcor and Hammond. I know the CXSE line of Edcor is $95, I don't know about Hammond.

2) Over $250 to $350: Sowter, Electra Print, Lundahl.

I want to hear from someone that use both the expensive and the cheap and tell me their impression. Is it really worth the extra money?


Also, if I just change the OPT in my cheap amp, would I expect to hear a difference ( of cause I mean improvement!!!)?


Also, if I buy new OPT, I would like to buy one that can use two tubes in parallel. The primary of the cheap transformer should be about 4K from calculation. If I buy one for two in parallel, I should buy one with 2K output impedance ( 2.5K is the closest I see). Would I degrade the sound if I put the new transformer into the cheap amp because it is miss matched?

For push-pull circuit, you may be satisfied with Hammond. But for SE, you need to get as best as you can - Lundahl is a decent brand.
 
Even if you could completely "fix" the amp you are still not ever going to get more than ~10W or so per channel. If that is a concern with your speakers then it's just good money after bad.

As far as doubling the outputs that will more or less double the load on the PT side, particularly if you stay with SE. It's certainly speculative if the transformer has the current on the HV side to do that and probably not on the filament side either.
 
It would definitely make a lot more sense to start from scratch than to replace the OPT on this amp, but the question remains if this is to work with your speakers or to learn about tube amps. If it's the former, either build PP or don't bother, you could maybe get away with 845 but that's not a beginner project (or something that could be done cheaply) and even then it's not that likely to satisfy. If it's the latter, buy or build a set of speakers with 96dB efficiency or higher, then you have a fighting chance with SE.
 
And, to that, rather than just buying new OPTs in hope of some revelation, perhaps use your measurement Kung fu to better understand the capabilities of the existing OPTs.
 
And, to that, rather than just buying new OPTs in hope of some revelation, perhaps use your measurement Kung fu to better understand the capabilities of the existing OPTs.

I suggest to find good SPICE simulation tool and model amplifier there first. From my experience 80% of circuit development can be done there before going to breadboard. There are multiple models of tubes available from various sources. For transformers most parameters can be measured and used in model.
 
Thanks guys
I am considering building an amp from ground scratch, I was just thinking it doesn't hurt to try the transformer in the cheap amp as it takes a while to build the new one, where it might only take a day to mod the transformer in the cheap amp and just listen. I am not paying big money to salvage the cheap amp, it's a goner already, the power transformer is not going to support it already.

Most important thing for me is to decide whether I want to pursue further. Money is not exactly a concern if I decide to pursue this. Ha ha, you don't want to know how much I spent on the two amps I designed and built!!! There's no cheap DIY, you want cheap, buy existing amps and be done with it!!!
 
I suggest to find good SPICE simulation tool and model amplifier there first. From my experience 80% of circuit development can be done there before going to breadboard. There are multiple models of tubes available from various sources. For transformers most parameters can be measured and used in model.
I use LTSpice, I have to find out whether I have model for EL34, I am sure I have 12AX7.

I used simulation when I designed the SS amps. I found they are not the most accurate in the world. I was quite good in simulation, I was talking to Yahoo LTSpice group on the forum doing the simulation at the time. I am sure the simulations were correct. I built the amp and did the measurement, it is very different. I went back and put in all the grounding, wire connections into the simulation. It was not close to the real world.

Regarding to SS, RF, grounding design, I have a lot of experience in pcb layout for signal integrity. I layout all my own pcb all these years at work that is 10 times more critical than amplifiers like this. I know I did not create problem with the layout. The real circuit does not behave the same consistently. Both stability and THD is not right from the simulation. I really would not hold my breath with the simulation. It can be used just for a checking major errors, not to be rely on as the end all means.
 
Alan0534...I think you are worrying too much about spec busting on tubes..If you built guitar amps you know that run the output tubes VERY hot.
Try visiting tubelab.com
 
Thanks guys
I am considering building an amp from ground scratch, I was just thinking it doesn't hurt to try the transformer in the cheap amp as it takes a while to build the new one, where it might only take a day to mod the transformer in the cheap amp and just listen. I am not paying big money to salvage the cheap amp, it's a goner already, the power transformer is not going to support it already.

Most important thing for me is to decide whether I want to pursue further. Money is not exactly a concern if I decide to pursue this. Ha ha, you don't want to know how much I spent on the two amps I designed and built!!! There's no cheap DIY, you want cheap, buy existing amps and be done with it!!!

You can start with something half-made. Like Yaqin MS-650B, or (better MS-845). Then re-work circuit to your satisfaction. Transformers and chassis is the majority of cost in these units.
 
I use LTSpice, I have to find out whether I have model for EL34, I am sure I have 12AX7.

I used simulation when I designed the SS amps. I found they are not the most accurate in the world. I was quite good in simulation, I was talking to Yahoo LTSpice group on the forum doing the simulation at the time. I am sure the simulations were correct. I built the amp and did the measurement, it is very different. I went back and put in all the grounding, wire connections into the simulation. It was not close to the real world.

Regarding to SS, RF, grounding design, I have a lot of experience in pcb layout for signal integrity. I layout all my own pcb all these years at work that is 10 times more critical than amplifiers like this. I know I did not create problem with the layout. The real circuit does not behave the same consistently. Both stability and THD is not right from the simulation. I really would not hold my breath with the simulation. It can be used just for a checking major errors, not to be rely on as the end all means.

Forget about grounding, signal integrity etc. When you need to simulate - basic operation with reasonably good tube and transformer models. Other components ( R & C ) can use simple models coming with tool. For transformer make sure model includes stray inductance and winding capacitance. Non-linear core magnetics are harder to simulate, but there are models for that too.

When basic circuit is ready, then you can build a prototype. There you start playing with layout for lowest noise, RFI etc.
 
Alan0534...I think you are worrying too much about spec busting on tubes..If you built guitar amps you know that run the output tubes VERY hot.
Try visiting tubelab.com
I get too used to SS design, you don't push, you don't want to come close, they burn!!! I never stop and think in the guitar amp, everyone use about 420V +B, there's no reason to worry. I actually lowered to 400 in my amp for the 6L6GC tube. High power is not a virtue for guitar amp, I want about 30W which is beyond 6V6GT ( 20W) but I don't want 40W ( normal for 6L6). So I cut the voltage!!! So bottom line, I never pushed my tubes before.
 
Forget about grounding, signal integrity etc. When you need to simulate - basic operation with reasonably good tube and transformer models. Other components ( R & C ) can use simple models coming with tool. For transformer make sure model includes stray inductance and winding capacitance. Non-linear core magnetics are harder to simulate, but there are models for that too.

When basic circuit is ready, then you can build a prototype. There you start playing with layout for lowest noise, RFI etc.

Here http://www.intusoft.com/articles/satcore.pdf is a reference to transformer models.
 
Thanks For_P1

I think LTSpice do have transformer model, I have to look it up, never done tube simulation before.

I think I need to study up tube design, I am using RDH4. It's going to be like a week or two.
 
Last edited:
But this is a popular kit, DIYAudio has a long thread on this A9 design.

Well, there are a lot of popular things that I'd argue aren't much good. I could name cars, bands, equipment, whatever.

The other thing to consider here is that numbers don't always tell you everything. I'm a fan of measurements, its the only truly objective way we have to compare stuff, but the fact is you can have a very pleasant sounding amp that measures pretty badly. A lot of people love Magnavox console amps, but in stock form quite frankly they're not awesome on paper. Conversely for years I hated the sound of my Phase Linear amp. Measures great, meets all its specs, but it made me want to plug my ears. I changed some parts and its much better now, still measures about the same but the sound is not the same. I can't quantify why its better, but it is.

Also, just because its a simple design doesn't mean there can't be "gotcha" elements to it. Layout, component quality, lots of other little things go into it. I've had far more problems getting basic 5 tube AM radios to work right than I have with multi-tube complicated shortwave receivers. The simple stuff sometimes has to be precisely just so in order to work, where the complicated stuff can be somewhat immune to less than perfect component placement and such.
 
Back
Top Bottom