How can I improve the bass response in my SET vacuum tube amplifier?
Hello instructables,
I'm experiencing an odd problem with my new SET amplifier. I just built Fred Nachbaur's MiniBlok 13EM7 SET amplifier and the bass response is poor with the power turned on, but when I turn off the plate supply and allow the amplifier to run on the energy stored in the capacitors, there is an immediate and very audible change in the rendering of the lower frequencies. I followed the schematic as closely as I could, but until my Edcor 10k:8 Ohm Output Transformer arrives, I am using a small 2K:8 Ohm Output Transformer from an old transistor amplifier. I have tried to install a filter choke and another capacitor in the power supply, but this did not improve the audio quality. Does anyone have any advice? Is the output transformer the problem, and if so, why?
Thank you!
I'm experiencing an odd problem with my new SET amplifier. I just built Fred Nachbaur's MiniBlok 13EM7 SET amplifier and the bass response is poor with the power turned on, but when I turn off the plate supply and allow the amplifier to run on the energy stored in the capacitors, there is an immediate and very audible change in the rendering of the lower frequencies. I followed the schematic as closely as I could, but until my Edcor 10k:8 Ohm Output Transformer arrives, I am using a small 2K:8 Ohm Output Transformer from an old transistor amplifier. I have tried to install a filter choke and another capacitor in the power supply, but this did not improve the audio quality. Does anyone have any advice? Is the output transformer the problem, and if so, why?
Thank you!
13
answers
|
Answer it!
|
sort by:
active |
newest |
oldest
May 8, 2010. 4:46 PMgmoon
says:
While many people tout the effects of using an undersized OT ("core saturation" :-P ) one thing you will NOT get with an undersized transformer is bass response.
Maybe the sound improves when you cut the power because the OT isn't over-saturating anymore.
And a 2K load resistance is WAY off what he's using in that schematic. Your new Edcor trannie is very close though:
35:1 winding = 1224:1 impedance, or 9.8K:8.
Assuming that's the root of your problem, the new transformer should make a big difference...
Maybe the sound improves when you cut the power because the OT isn't over-saturating anymore.
And a 2K load resistance is WAY off what he's using in that schematic. Your new Edcor trannie is very close though:
35:1 winding = 1224:1 impedance, or 9.8K:8.
Assuming that's the root of your problem, the new transformer should make a big difference...
May 8, 2010. 5:25 PMgmoon
says:
Hmmm. I just looked up the 13EM7.
The datasheet doesn't spec a load resistance. But using two ways I know of finding the correct load resistance:
1) For triodes, the OT Z is about 2-4 times the Rp (plate resistance) value.
The Rp is ~750 ohms @ 150V. Three times that is 2250 ohms.
(Unfortunately, the datasheet give Rp for only 150V.)
2) The other formula is Zout = Va/(Pa/Va), where
Va = plate voltage (150V)
Pa = max plate dissipation (10 watts)
150 / (10/150) = 2250 ohms. That follows--also for a plate voltage of 150V.
Of course, Fred's schematic shows 215V:
215 / (10/215) = 4622 ohms.
Perhaps he uses a higher load resistance because, as he writes:
In other words, even though we can get less usable power at higher load resistances, our linearity improves (and therefore, total distortion decreases).
(That's a blanket statement, and as such, untrue--there's certainly a point beyond which you shouldn't increase the load resistance.)
Regardless, if the OT is too small to handle 10 watts, it will saturate even if it's close to the correct impedance.
The datasheet doesn't spec a load resistance. But using two ways I know of finding the correct load resistance:
1) For triodes, the OT Z is about 2-4 times the Rp (plate resistance) value.
The Rp is ~750 ohms @ 150V. Three times that is 2250 ohms.
(Unfortunately, the datasheet give Rp for only 150V.)
2) The other formula is Zout = Va/(Pa/Va), where
Va = plate voltage (150V)
Pa = max plate dissipation (10 watts)
150 / (10/150) = 2250 ohms. That follows--also for a plate voltage of 150V.
Of course, Fred's schematic shows 215V:
215 / (10/215) = 4622 ohms.
Perhaps he uses a higher load resistance because, as he writes:
In other words, even though we can get less usable power at higher load resistances, our linearity improves (and therefore, total distortion decreases).
(That's a blanket statement, and as such, untrue--there's certainly a point beyond which you shouldn't increase the load resistance.)
Regardless, if the OT is too small to handle 10 watts, it will saturate even if it's close to the correct impedance.
gmoon - as it happens, I haven't yet ordered any transformers, I was about to but then decided not to until I found an answer. What output transformer would you recommend using? I am on a bit of a budget and the $35 Edcor was looking a bit too expensive if I were to go stereo.
May 8, 2010. 8:26 PMgmoon
says:
Now I see you're using a PP output transformer...that could be the problem--as well as the size.
It doesn't seem like you should get 200mA of current @215V. That's a lot of current through a tiny OT.
Also, a Pa of 10 watts is probably a max of 5 or 6 watts of output in a single-ended topology. Fred's only expecting 1 or 1.5 watts...that's probably why he chose a 9.8K OT, vs. a lower load resistance.
Is this an "audiophile" project? Because the choice of OT (and the output) will effect the level of acceptable distortion... I fool with guitar amps generally, which is a different thing. And he's not letting the signal clip, either (of course, once it clips it's not measured as "legitimate wattage" anyway, according to the definition.) Are you pushing it too hard for clean output?
AES has an 8 watt, 5K:8 single-ended transformer for $14. Search for P-T31(their new webpage is all frames, so I can't just copy the link..) Actually, I like Edcor transformers, too. They are HUGE, and I think over-engineered (the wattage specs are conservative.)
It doesn't seem like you should get 200mA of current @215V. That's a lot of current through a tiny OT.
Also, a Pa of 10 watts is probably a max of 5 or 6 watts of output in a single-ended topology. Fred's only expecting 1 or 1.5 watts...that's probably why he chose a 9.8K OT, vs. a lower load resistance.
Is this an "audiophile" project? Because the choice of OT (and the output) will effect the level of acceptable distortion... I fool with guitar amps generally, which is a different thing. And he's not letting the signal clip, either (of course, once it clips it's not measured as "legitimate wattage" anyway, according to the definition.) Are you pushing it too hard for clean output?
AES has an 8 watt, 5K:8 single-ended transformer for $14. Search for P-T31(their new webpage is all frames, so I can't just copy the link..) Actually, I like Edcor transformers, too. They are HUGE, and I think over-engineered (the wattage specs are conservative.)
While this isn't really an "audiophile" project (I don't have that kind of money, unfortunately), my goal is to make a solid tube amp (pardon the pun ;) which will replace my aging Sony SS System. This means that I won't want some of the effects associated with guitar amplifiers, and I might want to invest in some slightly nicer irons. However, I am planning a SE KT88 in the near future and wish to use this project to gain some skills with the basic metalworking involved in building the chassis without breaking the bank, so I do not want to spend more than about $100 at this point.
I do not believe that I am pushing the amp too far for clean output, there is an LED which indicates the charge in the capacitors, and it begins to pulse when the volume is turned all the way up, but not at lower levels.
I do not believe that I am pushing the amp too far for clean output, there is an LED which indicates the charge in the capacitors, and it begins to pulse when the volume is turned all the way up, but not at lower levels.
May 9, 2010. 6:31 AMgmoon
says:
When I built my SE amp I used a Hammond "universal" transformer; they're supposed to cover both SE and PP operation. Even though it was within the wattage specs, it was over-distorted--very little headroom.
I think both Steve and I are telling you the same thing--you can't use a PP transformer in a SE project.
Single-Ended amps conduct 100% of the current whether they are loud or silent. I.E., at quiescence or at maximum output the current through the transformer is the same. That's why SE are gapped, and always larger than PP transformers--there's a constant DC voltage on the transformer.
Before ripping out the transformer, you could add a NFB (negative feedback) loop, which will reduce distortion (it will reduce the output too, somewhat.) It's simple, and worth a try. Check out the Fender Champ 5E1, the NFB loop is the connection from the OT to the cathode of the second stage 12AX7--with the amount of feedback controlled by the 22K resistor (play with that value.)
One more thing--Fred's project is cool, but it's another one of those "oddball" tube projects. Lots to learn from it, but it's a compromise.
I think both Steve and I are telling you the same thing--you can't use a PP transformer in a SE project.
Single-Ended amps conduct 100% of the current whether they are loud or silent. I.E., at quiescence or at maximum output the current through the transformer is the same. That's why SE are gapped, and always larger than PP transformers--there's a constant DC voltage on the transformer.
Before ripping out the transformer, you could add a NFB (negative feedback) loop, which will reduce distortion (it will reduce the output too, somewhat.) It's simple, and worth a try. Check out the Fender Champ 5E1, the NFB loop is the connection from the OT to the cathode of the second stage 12AX7--with the amount of feedback controlled by the 22K resistor (play with that value.)
One more thing--Fred's project is cool, but it's another one of those "oddball" tube projects. Lots to learn from it, but it's a compromise.
Just out of curiosity, what sorts of effects are caused by impedance mismatching with the plate?
What quiescent current's flowing through the transformer ? I'd guess at a saturation effect.
Steve
Steve
I just measured 200mA through the transformer, and it's from a PP amplifier, so no air gap in the core. The thing measures about an inch tall by an inch and a half wide and half an inch deep - I think it MIGHT be saturating ;)
I just tried a larger 120V:6V power transformer (also no air gap) instead, but the sound quality only decreased. What sort of differences are there between impedance matching transformers and audio output transformers?
I think you need a gapped core. I also think that the core is VERY small for the power output....I shall consult an expert and get back to you....
May 8, 2010. 12:01 PMRe-design
says:
Can you give me a link to the schematic? Without that it's just a guess.
But if the output trans is wrong that might shift the freq. response.
Are there tone controls? If so you may have to adjust the caps there to shift the response to something you like more. That's done with guitar amps many times.
Also, what kind of speakers are you using?
But if the output trans is wrong that might shift the freq. response.
Are there tone controls? If so you may have to adjust the caps there to shift the response to something you like more. That's done with guitar amps many times.
Also, what kind of speakers are you using?
Here is a schematic: http://www.dogstar.dantimax.dk/tubestuf/miniblk2.htm
The actual output transformer does have the capability to reproduce the lower frequencies in question, but I suppose it's possible that impedance mismatching between the transformer and the triode's plate could cause this problem.
There are no tone controls in the amplifier, the component count in the signal path is minimal, why I chose this design.
The speakers I'm using are 8 Ohm impedance Yamaha NS-A835 models. They are a bit old, but by no means are they low quality speakers. I have read that they suffer from poor high frequency response, and that the tweeters were replaced in subsequent Yamaha models. I do not think that the frequency response problem has to do with my speakers, especially since the problem goes away when I turn off the plate supply.
Thank you for the help!
The actual output transformer does have the capability to reproduce the lower frequencies in question, but I suppose it's possible that impedance mismatching between the transformer and the triode's plate could cause this problem.
There are no tone controls in the amplifier, the component count in the signal path is minimal, why I chose this design.
The speakers I'm using are 8 Ohm impedance Yamaha NS-A835 models. They are a bit old, but by no means are they low quality speakers. I have read that they suffer from poor high frequency response, and that the tweeters were replaced in subsequent Yamaha models. I do not think that the frequency response problem has to do with my speakers, especially since the problem goes away when I turn off the plate supply.
Thank you for the help!
 |
