• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Adding NFB makes headphone amp sound AWFUL!!!

Status
Not open for further replies.
Jump to Latest
#21
A quote fest... sorry for that 😱

The original thread about this amp is here btw: http://www.diyaudio.com/forums/tube...mp-distorting-clipping-weak-power-supply.html

dludingt said:
That would lower the output impedance from about 280r to 40r.
Click to expand...
No way... as explained in the other thread, you probably had an output impedance around 3K to start with. With feedback, more like 150r.

On top of feedback, didn't you also add that 470uF cap at the cathode of the output tube ?

Bigun said:
There also needs to be enough current flowing through the FET so a power resistor from the FET source to ground maybe required if there isn't enough current being drawn through the output tube. Unfortunately, as I've never built one of these I have no experience to draw on. I'd be tempted to try and ensure enough current through the tube so I didn't have to add a power resistor.
Click to expand...
Sadly the power supply is very weak on that one as it was sized for a transformer coupled amp. One can easily run the whole current available through the output tube (about 20-25ma or so).

dirkwright said:
It looks like you don't have much of any options with this thing then, unless you want to buy 4:1 output transformers. Those will cost you at least $80 each.
Click to expand...
There's always those edcor which aren't bad at 11$ (for parafeed). Using the 7k and the 300r taps could work ok.


Btw, 10mv of offset with 300r sennheiser is really of little importance. I wouldn't suggest it for 16 or 32r headphones though, but this amp is very ill suited to such headphones anyway.
 
#22
Ben, how did you calculate those output impedances? I've done some research, and all of these ASL OTL modes run about 2-3Kr.

I did add the 470uf cathode bypass, which helped a bit.

Those Edcor OPT's look tempting. I've already worked out a tube rectified power supply that should have a B+ of about 275 VDC. I guess I could entirely rebuild this little thing at some point.
 
#23
dludingt said:
Ben, how did you calculate those output impedances? I've done some research, and all of these ASL OTL modes run about 2-3Kr.
Click to expand...
Since you originally had a simple triode gain stage, without a cap bypassing the cathode resistor and no feedback, the output impedance is the plate load in // with the internal resistance of the tube + the cathode resistor multiplied by µ+1. Doing it quickly, the plate load is very high (as a ccs) and thus doesn't matter, ra is 2k and µ 17. Thus 2000+ (18*75)= 3350. I rounded it down to 3K.

For the output with feedback, I'm lazy and relied on ltspice. You check the voltage output with 300r and 300k as load. The output impedance is calculated thus: (load resistance * unloaded voltage / loaded voltage) - load resistance
 
#24
00940 said:
Since you originally had a simple triode gain stage, without a cap bypassing the cathode resistor and no feedback, the output impedance is the plate load in // with the internal resistance of the tube + the cathode resistor multiplied by µ+1. Doing it quickly, the plate load is very high (as a ccs) and thus doesn't matter, ra is 2k and µ 17. Thus 2000+ (18*75)= 3350. I rounded it down to 3K.

For the output with feedback, I'm lazy and relied on ltspice. You check the voltage output with 300r and 300k as load. The output impedance is calculated thus: (load resistance * unloaded voltage / loaded voltage) - load resistance
Click to expand...

Ben,

Actually the cathode resistor is 150r, which gives me an output impedance closer to 6kr. Yikes. I think that the most cost-effective way forward would be first to get those OPTs in a parafeed configuration. Then a power supply could come later.
 
#25
Ignoring the qualities or weaknesses of the design, adding feedback can be problematic as Bigun noticed: there are ~2.5 phase lead circuits acting in the same frequency range: 2 in the couplings, and the rest in the cathode bypass.
That is not sufficient to turn the whole thing into an oscillator, because tubes have a low enough gain, but it probably creates a hump in the low frequency response. Adding another coupling cap would make things worse without solving any significant issue.
If feedback is desired at all costs, some kind of equalization or compensation would be required.
 
#27
Sure, there's better. Sowter has extremely nice transformers for headphones too.

But those little edcor are about the only cheap alternative you have. They're also used in the "torpedo", a design that received quite a lot of positive reviews.
 
#28
So, would I be able to put in nicer OT at a later time? If I put in OPT's, then there isn't too much I'd have to change at this time? Eventually, a new PS would be needed.

Thanks again everyone for the help. I've been taking a crash course in tube audio design but still feel like I am in over my head.
 
#29
Hi,

Wouldn't throw too much money at it. Enjoy it as it is, it wasn't meant to be high-end and you probably knew that from the get go anyway.

If it were me and given the quality of the headphones, I'd save up a bit and build a nice little amp/preamp for them when I'm good and ready.

There are plenty of good one out there and most of them can still be improved without breaking the bank.

Ciao, 😉
 
#30
That's probably good advice. When I mentioned this to my wife I got... The look.

I'm thinking that with the Edcor OPT with their low cost will make this a functional amp, even with the weak PS. Then, I can save up and build an entirely new amp when the funds are available.
 
Status
Not open for further replies.
Top Bottom