ECP Torpedo III [indexed in first post]

Yeah I didn't try R2 vs choke, either. It just seemed intuitive to eliminate ripple for the price of a cheap tube. I think you'd commented previously on a sonic difference, but had noted the Takman plate loading resistors were installed at the same time as the R2/choke swap. I may do some resistor swaps down the road, if boredom dictates. At that time, perhaps I'll also evaluate the choke vs R2.... but for now, I'm just enjoying the music.

 

I did report that, and the sonic difference was noticeable. But yes; I did both the electronic choke and the plate loading resistors at the same time, so I will reverse the mods to check where the influence come from. The plate loading resistors usually make a noticable difference since this is the most critical spot in the circuit for resistor sound quality. Plate loading resistors can make a difference in sound comparable to rolling tubes. So the question mark is more the electronic choke, than the resistors, especially in this circuit that is so optimized to have high CMRR. Reducing ripple should not really matter, but less ripple is always nice. And hearing is believing

I have ordered some more plate loading resistors and will make a temporary set-up so I can change them without having to pull the PCB out of the box. I will ''resistor roll'' the following:

- The standard resistors that comes with the kit
- Takman carbon (what I use now)
- TDK metal film
- Amtrans AMRG carbon

 

I like chokes in power supplies, but we have found with this form factor that they do more harm than good. Replacing R3 with a small CCS would probably do as much as the choke does, if not more as it will also lower the power supply's output impedance.

And the fact that this is an "electronic choke" should reduce the electrical noise considerably making placement less of an issue.

 

Since we are all cramming bigger and bigger caps in here, I think it is time to revisit capacitance size. We used the 10u caps because it ensured that frequency and phase issues remained well outside of hearing range. And I maintain that they are the correct choice.

However, relaxing that requirement may yield good results for some people. To wit, smaller caps will move a small bass bump up in frequency, which could warm the sound slightly, but it will also alter the phase a little which may or may not be audible. That said, the amount you can reduce the caps is larger than you might think. I think 5u caps will be indistinguishable for most people, and 2.2u caps will probably please midrange focused people.

The reason the graph mentions the mosfet output is that it explains the high frequency rolloff. Your amp is not doing that.

And apologies for the color blind - the bottom line on each graph is the 10u, the middle the 5u, and the top the 2.2u.

 

Yes please, that would be really interesting to test..! Lower impedance tend to make a positive sonic difference

Man Dsavitsk... You are opening a potentially expensive door with the talk about smaller caps

 

Hmm ... I've seen that 2.2uf response curve before.

 

Some nice resistors arrived...

The black with golden legs are the Amtrans 2W AMRG carbon resistors, and the brown ones are the TDK 2W CM2 metal film. Both types are quite a bit bigger than the standard plate loading resistors that comes with the kit - shown left. Also bigger, than the Takman carbon I use now. But it should be OK to make them fit...

Ready for some plate load resistor rolling

 

@MortenB how goes the resistor rolling?

 

It's really hot here these days, so we're sweating. Not the best for long listening sessions with headphones, so I have not started yet. Will do on the next rainy day

Why would I not like the Amtrans?

 

Just a guess based on your stated preferences, but let's not bias the experiment... More curious than anything, so I hope it rains there soon.

 

Hmmm... I don't get this. When reading your comments on ''my stated preference'' I get the impression, that your perception of my preference is very different from what my preference actually is. Discussing something as subjective as sound, with people we don't know, is very difficult. But here is what I'm looking for with this resistor rolling:

1.
I'm trying to learn the sonic signature of the circuit by small experiments, that are known to have an effect on sound - the plate loading resistors is a good example.

2.
I'm trying to balance the sound to become what I find natural, and this is where I don't get your comment on ''stated preference''. To me the sound of the T3 is a bit lean. I think this is because of the Mundorf caps. They are an amazing find for the T3 since the form / factor seems like they were made for the T3. So a big thank you to the one who found them..! These Mundorfs are known to be a bit lean sounding though. It seems that many here in the thread balance this out by using full and warm sounding tubes from Mullard and Brimar. And so do I...

My preference is not for a lean sound, my preference is for a natural sound. So what is that? Don't remember if I have mentioned this before, but I'm a musician so I get to hear how instruments and voices sound live on a daily/weekly basis. So when I say natural sound it's the sound of instruments and voices ''live''. I know that our recordings are mastered, edited and what not, but I still aim for a natural sound with reference to how instruments and voices sound in the real world. And with this as a reference point I find the T3 slightly lean when using neutral sounding tubes. I find the T3 more natural sounding when using the British tubes with more meat on the bones.

So actually, I think that I will prefer the Amtrans since these resistors are know to have a warm sound and a treble that is a bit mellow (a bit of an old school carbon resistor sound, but with good transparency). I'm hoping they will balance out the Mundorf caps. The TDK resistors should be the opposite: Very clear and neutral, and the Takman that I use now somewhere in between. But hearing is believing, so time and listening will tell. Maybe it sound totally different...

I have used the Takmans before, so I know them to have a quite neutral sound with just a hint of warmth. I have not tried the Amtrans, but they are developed as a replacement for the discontinued Riken Ohm, and I have used those before. So it will be interesting to check if the Amtrans are similar. The TDK's I have never tried.

 

Ah, interesting read... And I'm very curious to what extent resistors can influence the sound to the degree you're hoping for. I had recalled in the past that Kiwames were not to your liking, and have picked up the impression elsewhere in DIY forums that the Amtrans carbons were similar in character. That's what I was basing it on. But now that you've had time with the T3 and have a feel for the direction you'd like to take it, perhaps either or both will do the job!

I'm also considering the yellow Brimars now since they seem to be so popular. I just don't want to give up much in the way of soundstage to other tubes like the more neutral 7062 which does have a more expansive stage than say my Mullards... I don't want to give up much, because I feel the T3's stage is bit compressed already. So maybe resistors can have that same warming effect without the tube swap. Interested to find it.

 

I have tried Kiwame carbon resistors in both speaker cross overs and in tube amps, but I always end up not using them. I feel they lack transparency and this gives kind of a muddy sound (English second language and I can't find a better word). Other resistors have meat on the bones without the loss of transparency. Shinkoh tantalums, Riken Ohm carbon etc. If the Amtrans also lack transparency, then I don't think I will use them - you're right on that.

I don't know this circuit well (yet), but in other tube amps I have heard similar effects on sound with different plate loading resistors as when swapping out tubes or caps. The plate loading resistor is THE place in a circuit where resistors make the most difference in sound. If there is a feedback resistor it also has big effect, but we don't have that here. The full current of the tube circuit in the T3 runs through the plate loading resistor and the sonic signature is bigger where current is bigger.

Resistors are much cheaper than tubes and parafeed caps, so IMHO it's an obvious place for experiments. Maybe the plate loading resistors don't have that much effect on the sound in the T3, but they are cheap, and I will have learned this about the circuit. Maybe they do have a good effect, and I will have balanced the Mundorfs for just a few bucks...

 

Has anyone tried the Psvane 12AT7 tubes in the T3..? I would love to hear your impressions...
http://psvanetube.com/wordpress/store/products/12at7-tii-pair-or-single/

 

Interesting I might try the 12au7 one.

 

Hmmm... Do you think you will have enough gain with the 12au7? Check this, if you have not seen it already:
http://www.diyforums.org/Torpedo-III/TorpedoIII-tubes.php

 

For the money, I think I would pick this over the Psvanes. My experience is that the Psvanes are good, not great.

http://www.upscaleaudio.com/12at7-ecc81-6201-cv4024/mullard-cv4024-12at7/

 

Yeah it's enough gain for HD650, I have a GE 5 star 12au7 and it has a enough gain. I like having the pot near max because it should lower the SNR since noise is constant while the input signal is stronger.

 

Usually, but probably not here. The circuit itself is very quiet. The only appreciable noise is what is picked up on the input line. That means that for a given overall gain, the noise on the output will be the same regardless of tube (with the exception that tubes with lower heater draw will cause less stress on the power transformer which will in turn lead to less noise being picked up.) That is, if you want to listen at volume XdB, whether you use a low gain tube and turn the volume up, or a high gain tube and do not, the noise will be exactly the same. The only way to defeat this is with a higher signal level on the input.

 

Thanks for clarifying, I just assumed it was best using the lowest gain as it produces the lowest amount of noise.

Wouldn't turning up the pot to maximum = higher signal level on the input?

edit: nevermind I get what you mean turning up the pot also raises the noise from the input as well.

But also pots sounds better maxed out and the input pot goes before the differential part of the circuit.