• 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.

Eico HF-81 C23 C24 Value change

Jump to Latest
R

rjb5191

Member
Joined 2022
#1
Hello all, It's my first post here!

C23 and C24 in the Eico HF-81 in question here. Schematic found here: https://www.thetubestore.com/lib/thetubestore/schematics/Eico/Eico-HF-81-Owner-Manual-Schematic.pdf

It looks like these are local negative feedback coupling caps (150pf). One of these is physically damaged and I would like to replace them. The issue is that I don't have the proper value and voltage rating on hand but I do have 430 pf 630v polystyrene caps on hand so I'm wondering how this will impact the performance of this stage. I'm getting different results from online calculators for this circuit. One is showing me it's essentially flat with the 150 pf cap and the other shows that the -3 dB is around 120 Hz. I'm assuming a load impedance of 5M for the grid of the 12ax7 (V5b, V6b). If the circuit is already fairly flat, then I'd assume raising the cap value shouldn't effect the response in a way to be sonically noticeable or detrimental in any way.

Thanks for any help!
 
#3
Those caps are used to roll off the high frequencies in the forward path and don't have any real effect at 120Hz. This is done for stability with feedback around the whole amplifier. You really don't want to change their value without a very compelling reason. 150pF 630V styrenes shound be fine.

All good fortune,
Chris
 
  • Like
Reactions: jan.didden
#4
Chris Hornbeck said:
Those caps are used to roll off the high frequencies in the forward path and don't have any real effect at 120Hz. This is done for stability with feedback around the whole amplifier. You really don't want to change their value without a very compelling reason. 150pF 630V styrenes shound be fine.

All good fortune,
Chris
Click to expand...
Are you sure about that? That circuit doesn't look like a low pass filter to me. Please see the difference between the two schematics; Let's work off the schematic from my very first post. If you said C25-28, I would agree with you. Those are definitely part of a LPF, rolling off the highs.

I see a local feedback loop from the plate of V5b through R45 and R41, then through c23 back to the grid of V5b.
 
#5
Yeah, those are also low pass filters, but from such a low source impedance as to contribute only periferably even somewhat above the audio range, except to reduce slew rate. The C23, C24 are the fundamental "dominant pole" rolloffs, with resistive gain stops. You really don't want to change their value without a very good reason, but you might experiment with removing C25-C28, maybe add an output Zobel, if you really want to experiment.

All good fortune,
Chris
 
#6
I actually already lowered th value of c25-c28 from 220 pf to 150pf which seemed to open up the top end a bit (sorry no, measurements). Going by the terminology your are using, I'm guessing the dominant pole rolloff is to ensure stability and prevent high frequency oscillation. If you have the time, do you mind providing a calculation or the equations I need to use for this particular circuit to calculate that rolloff? I think that Eico was being conservative in this circuit due to the variability in kit builds so may have built in a large margin for stability. It's not that I feel I need to change it but I want to know where we are at here. It honestly sounds pretty good at this point so I'm fine just substituting in higher quality c0g caps of the same value.
 
#12
rjb5191 said:
a calculation or the equations I need to use for this particular circuit to calculate that rolloff?
Click to expand...
It's possible to calculate compensation values (these days most folk would use LTSpice) but that requires your knowing details about the output transformer and load, and making some big-ish assumptions that lumped-sum values apply to a distributed-values reality (leakage inductance and various capacitances). Your goal is to have a margin of safety above the point of oscillation, where gain is unity or greater when phase shift passes 180 degrees, measured as input vs. fed-back signal all the way around the loop back to the input triode's cathode, with the feedback disconnected and a dummy 1K8 resistor loading it.

Measuring your real amplifier, with a decent assumption of a real load, you'll observe relative magnitude (voltage) and phase (timing) and plot them together on a common x-axis (frequency). Congratulations, you've made a fancy Bode plot. Adjust the forward and reverse (feedback) path lag and lead for a good margin of stability. Done.

All good fortune,
Chris

ps: do yourself a favor and add a Zobel across the output. Maybe 0u1F in series with 10R across 8 Ohms or twice that Z across the whole secondary.
 
Last edited:
#14
All feedback amplifiers need a Zobel and a small isolating inductor for best performance when connected to the cold, cruel, somewhat capacitive outside world. Their stability-mandated frequency response looks inductive viewed from the outside, and tries to resonate with their ultimately capacitive external load. "Toob" guys classically try to ignore this, but that doesn't make it go away.

FWIW, amplifier inputs fed by an interconnect cable also benefit from a Zobel of the characteristic impedance of the cable. Unterminated cables have all kinds of weird response anomalies just above the audible range. Of course nobody bothers, because that's not how it was done in 1962, or in Japan in 1987, or in NYC in 1994, by gurus.

All good fortune,
Chris
 
You must log in or register to reply here.
Top Bottom