Eico HF-20 THD Evaluation

[Walter R]

My bedroom stereo has two monaural integrated Williamson amplifiers. One is an Eico HF-20 and the other was made by Elmat but appears to be an HF-20 clone. The Elmat has the same tube lineup and controls as the Eico but the power transformer appears different. I was curious about the THD of these amplifiers. The Eico manual only says “Mid-band - 0.3%”. Because my usual source levels are low, I installed a Douk Audio stereo preamplifier to feed the audio into the aux input of the amplifiers. The preamplifier has two 12AX7 and two 12AU7 tubes. The specifications for the preamplifier claim “THD<0.03%”.

I used old school instruments to measure the THD. That included an HP 334A distortion analyzer and a General Radio 1310a signal generator. The specs for the signal generator promise THD of about 0.1%. The amp under evaluation was driving an 8 ohm 20 watt non-inductive resistor. The signal generator was held at 2 volts rms output during all tests at all frequencies, from 20 Hz to 20 kHz. The amplifier volume was at 8.5/10, the base and treble controls were near neutral, and the loudness control was at 6/10. The preamplifier level control was at about 6/10 and was not changed during the testing. There wasn’t much voltage gain in the preamp during testing.

The attached chart shows the resulting THD values against a logarithmic frequency scale. First I measured the THD of the signal generator output. It was indeed close to 0.1% for the entire frequency range. Next I measured the THD of the preamplifier output. Within experimental error it was virtually identical to the signal generator THD, so that the 0.03% spec value is believable.

Next I measured the THD of the Eico amplifier output. If we subtract the preamplifier THD from the Eico values the results came out quite good, less than the specified 0.3% midrange and about 0.5% at the high and low frequencies.

Finally I measured the THD of the Elmat amplifier. It was very high, even if we throw out the large peak at 2 kHz as an experimental fluke. Either the Elmat is a dog or something is wrong with it. Both amps have relatively new J/J tubes in them. I have a third, spare amplifier, another Eico HF-20. My next step will be to substitute it for the Elmat and to see if the good HF-20 performance holds up.

I invite your comments and suggestions.

 

[gadget73]

try swapping the tube set in the HF-20 with the Elmat amp. Just curious if its as simple as a junk tube. Failing that I think you'd have to do a more detailed analysis comparing component values between the two amps to see what, if anything is different as well as confirming all the components and voltages are in tolerance.

THD figures on the Eico look entirely reasonable though.

 

[GordonW]

That big peak at 2K smacks of a possible noise issue (noisy tube or resistor or such)...

Regards,
Gordon.

 

[PakProtector]

The 32004 output in the HF20/22 is not to be despised... Thanks for the measurements.
cheers,
Douglas

 

[dcgillespie]

The one pic of an Elmat model WAK amplifier I could find shows it to be so much of a clone of the HF-20, that in all likelihood Eico even provided the chassis for them. Importantly however, the output transformer appeared to be at least physically the same in all respects as that of the original Eico piece, and undoubtedly is the same piece. If the output transformer of your Elmat amplifier appears to be different from that on your Eico (assuming it is an original piece), then that may not be the issue, but that detail certainly needs to be chased down to see exactly what difference it might be bringing to the table.

Dave

 

[Walter R]

The one pic of an Elmat model WAK amplifier I could find shows it to be so much of a clone of the HF-20, that in all likelihood Eico even provided the chassis for them. Importantly however, the output transformer appeared to be at least physically the same in all respects as that of the original Eico piece, and undoubtedly is the same piece. If the output transformer of your Elmat amplifier appears to be different from that on your Eico (assuming it is an original piece), then that may not be the issue, but that detail certainly needs to be chased down to see exactly what difference it might be bringing to the table.

Dave

I misspoke. It’s the power transformer that appears a little different. Thanks for the model number.
Walt

 

[Walter R]

try swapping the tube set in the HF-20 with the Elmat amp. Just curious if its as simple as a junk tube. Failing that I think you'd have to do a more detailed analysis comparing component values between the two amps to see what, if anything is different as well as confirming all the components and voltages are in tolerance.

THD figures on the Eico look entirely reasonable though.

I’ll try tube replacement this afternoon to see if that helps. Thanks!
Walt

 

[Walter R]

That big peak at 2K smacks of a possible noise issue (noisy tube or resistor or such)...

Regards,
Gordon.

I’ll start with tube replacement. Thanks

 

[Walter R]

I found a bad Chinese 12AX7 or 12AU7 in the Douk preamplifier that was causing the spike at 2kHz in the Elmat THD results. I hadn’t checked the THD in both channels of the preamplifier. I replaced the tubes with new J/J tubes. The Elmat THD values are now more reasonable (see revised chart) but are still very high above 2kHz. I’ll continue by replacing tubes in the Elmat to see if I can reduce it. None of the replaced tubes checked out as bad on my TV-7.

 

[Walter R]

My next step was to replace all the tubes in the Elmat amplifier with my new, spare J/J tubes, including a matched pair of 6L6GC. This improved the THD, as seen in the attached revised chart. But the Elmat THD remains 0.4% to 0.6% higher than the Eico THD across the frequency range. I subsequently tested the Elmat 6L6GC tubes on my TV-7 and one of them tested unacceptable. In push pull operation that could introduce a lot of distortion. This is about as far as I go in checking my amp performance, unless someone has further suggestions.

I started the distortion checking as a technical exercise, due to curiosity more than anything else. It turned out as an education in the importance of tube selection.

 

[gadget73]

Just to rule out variables, have you tried the tube set from the Eico to see how the Elmat does with a known good tube set? Or at least I'd assume they are good since the distortion numbers come out normally.

 

[zackthedog]

My next step was to replace all the tubes in the Elmat amplifier with my new, spare J/J tubes, including a matched pair of 6L6GC. This improved the THD, as seen in the attached revised chart. But the Elmat THD remains 0.4% to 0.6% higher than the Eico THD across the frequency range. I subsequently tested the Elmat 6L6GC tubes on my TV-7 and one of them tested unacceptable. In push pull operation that could introduce a lot of distortion. This is about as far as I go in checking my amp performance, unless someone has further suggestions.

I started the distortion checking as a technical exercise, due to curiosity more than anything else. It turned out as an education in the importance of tube selection.

Do you know how the Elmat circuit compares to the Eico circuit? Are they the identical? Also, while the OPT *looks* the same as the Eico's it's possible it was wound to less demanding specs, or maybe Elmat picked up seconds rejected by Eico? ;-) At any rate, I've got a pair of those 30024's in the shed, I should get busy and build something with them.!

 

[Walter R]

Do you know how the Elmat circuit compares to the Eico circuit? Are they the identical? Also, while the OPT *looks* the same as the Eico's it's possible it was wound to less demanding specs, or maybe Elmat picked up seconds rejected by Eico? ;-) At any rate, I've got a pair of those 30024's in the shed, I should get busy and build something with them.!

I have no schematic or other information on the Elmat amp.

 

[Walter R]

Just to rule out variables, have you tried the tube set from the Eico to see how the Elmat does with a known good tube set? Or at least I'd assume they are good since the distortion numbers come out normally.

I just finished doing that, it made no real difference in the Elmat THD.

 

[gadget73]

ok so we know its something within the amplifier that differs. Honestly at this point you'd have to start comparing the two amps for circuit and component values. If they use the same output transformer, seems reasonable that you could just modify the Elmat to have the same values as the Eico.

 

[zackthedog]

ok so we know its something within the amplifier that differs. Honestly at this point you'd have to start comparing the two amps for circuit and component values. If they use the same output transformer, seems reasonable that you could just modify the Elmat to have the same values as the Eico.

Yes, as gadget says, take the HF-20 schematic and compare it to the Elmat internals. Seems like you could "refurbish" the Elmat to match the Eico.

 

[gadget73]

would also be worth confirming measured vs marked values. I've seen old resistors that were completely out to lunch.

 

[Walter R]

Because my HP 334A distortion analyzer and General Radio 1310A signal generator are quite old, I felt I needed to check their calibration, I bought a Chinese Longwei TDM-1911 distortion meter to do this. The TDM-1911 is rather limited in that it only checks distortion at 400 Hz and 1000 Hz, but it’s new and calibrated. The resulting tests agreed well with my original results. The (skimpy) manual for the TDM-1911 isn’t available online, but I got one with the instrument. It’s attached.