Fisher 500B L-R frequency response differences

[charlesaudio]

Hi,

Any cue where to look at for differences in L-R frequency response for a 500-B ? Please see the included curves (division are 0.5 dB). Something might be wrong with one of the PEC. I would like to bypass to the tone stack but I don't know how to do it while still keep same gain, impedance etc.

Any help would be appreciated.

Thanks

 

[walyfd]

Can't help you except to say I have the same problem as you... right channel is always 1 to 2 volume control notches higher than the left...

 

[Tim D]

Does it happen on all sources? Have you tried swapping the tubes between channels? Has any work been done to it? Have you checked the voltages throughout and compared it with the voltages on the schematic? Have you had the tubes checked? You might find one or more tubes that measure distinctly different than the others. Have you cleaned the controls and switches?

I had an issue with my KX-100 amp being out of balance and ended up being a bad output transformer and a few out of spec and bad parts. I ended up measuring all the resistors and replacing a few. I ended up replacing the coupling caps. Part of my issue ended up being a bad ceramic cap. Not easy to find.

Dave and Larry and others may come along and give better and more detailed suggestions, but these are some things to consider.

 

[charlesaudio]

Yes I should have been more precise :
It the same on aux and tape-mon input, I did swapped the tubes with no change.
I recently replaced the power switch, put a cl-80, replaced a few coupling and psu cap, replaced grid res with 220k, 10ohm on pin5-0V, and built Dave's bias-bal mod.

The signal is the same amplitude on the reverb loop just before the tone control. It's really at the output of the next tube that it lack some mid and high frequency on the left side compared to the right.

 

[dcgillespie]

With the tone controls set at flat, check the response at the output of PC4 (terminal 4). Check it in both channels to check for equal results. Don't get too concerned if the results are not ideal at this point, as your test equipment will likely load down this high impedance point in the circuit to produce less than optimum results. The main thing is to see if the measured response is the same in both channels at this point. If it is, then with one tone stack being good, and identical response measured at the same point in the other channel, then that eliminates the tone stack in the bad channel as a possibility.

If the response is good at that point, then C39 or PC-6 may be suffering in the bad channel. Failure or well off-value components here could easily cause reduced HF response.

Let us know!

Dave

 

[charlesaudio]

Well, there is a minor difference at the output of PC4 but I think it's marginal at about 0,5dB difference. The difference is also marginal at the output of the tone stack tube so I was probably wrong saying it was ''no good'' there.

I was later probing with my scope on the speaker output while playing a 10k and it wasn't marginal there. I played with the speaker phase reverse switch and got more signal on the weak output. I sprayed some deoxit there and it got better.

One thing for sure is that there is some evident L-R balance variation and frequency response variation between L-R depending on the volume position. That's quite normal for balance but I'm not sure for frequency response.

According to the owner's manual, the overall frequency response should be within +/-1.5dB from 20hz to 25khz. The 500-B I'm working on doesn't meet that specification.

0,5dB/div


edit : looking at those curves in the hf region, it might reflect the negligible disparity I measured at the output of the tone stack. I removed c39 (it's not inside PC6) and it did not change anything.

 

[dcgillespie]

Charles -- Two points to consider:

1. Removing C39 may not have made a difference in the audio band, but it should make quite a difference in the supersonic band, where audio transients lie and are either passed, or not. C39 and R52 -- in addition to the 3 internal caps and two resistors of PC-6, make up a HF compensation network that counters Miller in the triode section of V10 that PC-6 works with when the High Filter switch is turned off. The result is basically flat response in the audio band. When the High Filter is turned on, then the same components form a sharp low pass filter to attenuate high frequencies accordingly.

2. Often, if test conditions are not precisely followed, it can lead to inaccurate conclusions. It is customary to measure frequency response at an output level of 1.0 watt RMS, with all the controls and switches set for flat operation, and the volume control advanced to the full up position. This latter point is important so that Miller in the AF Amplifier section of the driver tubes produces the optimum amount of HF roll-off for the power amplifier section. It is common for designs such as this employing a high resistance volume control driving a Hi Mu triode stage to show significant HF roll-off when the volume control is positioned at some mid point setting.

I hope this helps!

Dave

 

[charlesaudio]

Thank you for your time Dave, I really appreciate your input on this.

I will test it at 1W with volume ctrl at full up position. Last plot I posted is 1W with volume ctrl at around 10:00.

I'm sure C39 is there for something, I meant that by lifting it, the signal (10khz) stayed at the same level while observing the oscilloscope.

 

[dcgillespie]

I'm sure that a 10 kHz sine wave likely did show little change when C39 was lifted. However, if the signal is changed to a 10 kHz square wave, there should be a significant change to the leading edge of the waveform as C39 is connected or removed from the circuit.

Square waves are a wonderful tool for making very quick and easy response tests, as to pass a square wave faithfully, the circuit must have a response the is at least equal to the fundamental /10, up to the fundamental X 10. Therefore, to produce a 10 kHz square wave accurately, the circuit must have a response from 1 kHz up to 100 kHz. This upper band requirement is unreasonable for most circuits containing frequency sensitive controls, so when testing the response of tone control circuits, etc, it is best to use 200 Hz and 2 kHz square waves, which will then cover the full 20 Hz to 20 kHz band pass. Square waves also make it very easy to see the effects that the various controls and filters have on the waveform presentation.

Dave

 

[Dave451]

One thing Dave taught me which I have used to good effect is to use the scope with a square wave input into the amp to center the tone controls; not just rely on physical centering of the knob. With a 2 kHz square wave through the front end, it's easy to see when the tone controls are in 'neutral.'

 

[gadget73]

agreed on using other means to center the controls. I've found some have a significant cut or boost when set to "flat", either because the control is off-value or the circuit was designed to add some flavor to the music. I correct them where I can, or just make note of where flat frequency response actually is if its not a spline shaft that allows for easy repositioning of the knob.