The “Felpham” valve tester measurement circuit and pcb

[Electronpusher0]

Originally Posted by Diabolical Artificer

Questions, I couldn't find R31. Are all the RV's 10T presets? Lastly why all the paralleled 47n caps on the power supply filtering instead of big electrolytics?

Andy.

Hi Andy,
R31 is connected to the output of U3A (pin 1), its part of the 20Hz filter.

Yes the presets are multi-turn presets.

Ed has answered the question about the 47nF caps. The actual value is not critical, 0.1uF are more commonly used but I have a large box of 47nF!

I attach the V1.2 Schematic, this has the components mentioned in the Errata included (R45, C27)

Peter

[Diabolical Artificer]

Thanks Peter. Andy.

[Pamphonica]

Originally Posted by Diabolical Artificer

Ta, missed that. Looking at the PCB each IC has a cap across the rails.

Andy.

Hi Andy,
I've just checked on the v 1.2 schematic. Is that the master document?

If so, could you update the BOM for future reference?
If the schematic is correct then I think the updates are:


R1 - 100R (doesn't need to be 1% as swamped by the 390K string)

R8 - 5M6 (no note)
R11 - 9K1 not 91K
R13 - 9K1 not 91K
R24 - 8K2 not 56K

R31 - 82K
R32 - 5K6 not 4K7
R34 - 82K not 8K2
R37 - 10K not 4K7


C2, C3 - 10u/25V
C17 - 0.1u (shows as 1u on my copy)
C21 - 0.1u (shows as 1u on my copy)
C23 - 1u


RV1, RV2 - 2K not 20K
RV5 - 10K not 5K


All RVs 10 turn preset


I'm looking forward to trying this all out - I am now the proud posessor of a Mk1 pcb!


-Jeremy

[Diabolical Artificer]

Thanks Jeremy, will do. bit busy at present will do so ASAP.

Andy.

[Electronpusher0]

This needs a bit of explanation.

There are several differences between V1 and V1.2
I last worked on this project a couple of years ago and forgot that I had made these changes. Below is an explanation which will address the anomalies Jeremy has identified.

Oscillator
R32 was a 4k7 on V1. I have found that a parallel combination of 5k6 and 39k gives a better sine wave. 4k7 does work however.
Oscillator Output.
R37 and RV5 set the oscillator output. The actual value is not particularly critical so long as R37 and RV5 have the approximately the same value.
So RV5 as 10k requires R37 to be 10K
RV5 as 5K requires R37 to be 5k1
Output voltage scaling.
The output voltage for the anode and screen volts is select-able when assembling the board.
If a scaling of 1V per 100V of HT is required then fit the following values
R11 and R13 - 91k, RV1 and RV2 - 20k, R24 - 56k
If a scaling of 0.1V (100mV) per 100V of HT is required then fit the following values
R11 and R13 - 91k, RV1 and RV2 - 20k, R24 - 8k2

R1 compensates for the 100ohm measurement resistor and improves the common mode rejection so 1% is not needed.
low pass 20Hz filter
V1 used 1uF capacitors, I realized that 0.1uF caps are easier to obtain and changed R31 and R32 accordingly.
If using 1uF caps for C17 and C21 then R31 and R34 are 8k2
If using 0.1uF caps for C17 and C21 then R31 and R34 are 82k

Peter

[John_BS]

A few observations:


If R45 is fitted, C21 should be 1uF

There's a minor typo in the text above U1b - should read link R10, not R11
Surely R1 is redundant unless you've managed to match the 390k chains to 0.01%?

John

[Electronpusher0]

Originally Posted by John_BS

A few observations:


If R45 is fitted, C21 should be 1uF

There's a minor typo in the text above U1b - should read link R10, not R11
Surely R1 is redundant unless you've managed to match the 390k chains to 0.01%?

John

re C21 I will have to check on LTspice.

The typo is a good catch, thanks, should read as John suggests

R1, I probably did match the 390K chains to close to 0.01%. I bought 100 1% on a bandolier so they were from the same batch and then measured them all. I used the closest matching 12 for the board.

Peter

[John_BS]

Peter: I should have put a question mark after the 1uF, as the filter will probably still be out of kilter with R45 fitted, as R34 remains 82k. But it should be closer than leaving C2 at 0.1uF. Spice should reveal any compromise.

John