23rd Feb 2014, 8:30 am | #1101 |
Rest in Peace
Join Date: Mar 2011
Location: Basildon, Essex, UK.
Posts: 4,100
|
Re: The "Sussex" Homebrew Valve Tester.
I have just cut one leg of D16. and guess what, hooray it WORKS.
Very special thanks to Howard "Omegaman" and thanks to everyone else who has tried to help me. I have just got to put things back together now. It would be worth the others who were experiencing similar problems to me trying the same, remove the protection diode D16 Would a blocking capacitor help protect the GM meter from overload in the event of a flash over in a faulty valve. Mike |
23rd Feb 2014, 12:49 pm | #1102 |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: The "Sussex" Homebrew Valve Tester.
You could always put a switch in series with the diode, if it only causes problems with certain valves.
According to a certain well-known law, the additional inductance of the switch leads will counteract the phenomenon responsible for the misbehaviour, even when the switch is "on" .....
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
23rd Feb 2014, 1:39 pm | #1103 |
Hexode
Join Date: Dec 2009
Location: Leeds, West Yorkshire, UK.
Posts: 250
|
Re: The "Sussex" Homebrew Valve Tester.
Mike,
I'm glad that helped. It had me scratching my head for a while. All I need to do now is re-loom my wiring and tidy everything back up! I can then post some details of my interpretation of the design. I did my own pcb layout to incorporate all the relays and components previously mounted off the board. It may be of interest to others. Regards all
__________________
Howard G7AJN/M3OCL "How hard can it be?" - Jeremy Clarkson |
23rd Feb 2014, 4:03 pm | #1104 |
Octode
Join Date: Mar 2006
Location: Watford, Hertfordshire, UK.
Posts: 1,270
|
Re: The "Sussex" Homebrew Valve Tester.
That's very interesting as in theory the diode was expected not to influence the 200mV maximum expected across the sense resistor feeding the gm panel meter. I apologise to everyone who may have experienced a problem with this diode, it all seemed OK on my Sussex and so far I have not had to replace the gm meter. I had previously tried capacitive coupling the panel meter and other kinds of measures but none seemed to work. Perhaps I put too much faith in such a simple yet seemingly effective way of protecting the meter. In the meantime, I shall try to find a large bush to hide behind
__________________
Whether the Top Cap is Grid or Anode - touching it will give you a buzz either way! |
23rd Feb 2014, 6:00 pm | #1105 |
Hexode
Join Date: Dec 2009
Location: Leeds, West Yorkshire, UK.
Posts: 250
|
Re: The "Sussex" Homebrew Valve Tester.
Les,
No need to hide. The first Sussex I built was destroyed when a 6L6 flashed over. The second was built with the diode and so far the meters have survived a couple of flashovers. The third one didn't like the diode and sulked.... Regards
__________________
Howard G7AJN/M3OCL "How hard can it be?" - Jeremy Clarkson |
23rd Feb 2014, 6:09 pm | #1106 |
Octode
Join Date: May 2009
Location: Linkoping, Sweden
Posts: 1,465
|
Re: The "Sussex" Homebrew Valve Tester.
Howard, do you know if you changed any of the components between your three editions? Maybe different mosfets, diodes or transistors or meters? It would perhaps shed some light on why some editions work and some don't.
/Martin
__________________
Martin, Sweden |
23rd Feb 2014, 6:42 pm | #1107 |
Rest in Peace
Join Date: Mar 2011
Location: Basildon, Essex, UK.
Posts: 4,100
|
Re: The "Sussex" Homebrew Valve Tester.
Hi Les
Dont worry it looks like it is a random thing where this seemingly innocent diode upsets things. The scope traces that I have done since cutting the diode all seem similar to as they were before. So they are possibly fairly normal. It seems that there is a 50Hz pulse that builds up at the point of maximum conductance and lower bias, at this point the trace goes haywire, but the valve is still conducting and the GM is still OK so long as you dont wind the bias down too much further. It would still be interesting to me if I could compare the trace of my anode, screen and grid supply with some others at the correct grid setting, slightly more negative and also slightly negative. What happens during a "flash over", surely it can only be high voltage DC that can appear at this point.? If the meter is protected from DC by an on board capacitor it may only be a few 10's of volts and if high voltage DC is suddenly applied to the meter this in itself may flash over delivering HT DC to the meter electronics. Would a 630v .1uf cap block these higher voltages safely and be sufficiently "invisible" to the AC voltages. Mike |
23rd Feb 2014, 7:23 pm | #1108 |
Octode
Join Date: Mar 2006
Location: Watford, Hertfordshire, UK.
Posts: 1,270
|
Re: The "Sussex" Homebrew Valve Tester.
Hi Mike, I think the problem may be due to the much higher anode currents when testing things like EL34's. The DC voltage across the 10 ohm resistor would be sufficient to make the diode conduct whereas on small signal valves this problem never occurs, just guessing but I think quite probable. I suppose one could try 2 diodes in series but I often wonder if something like a 1:1 transformer would be better for sampling. It would have to have a good air gap to stop saturation as in the case of a single ended output transformer. Would a power transformer be up to the job? Which makes me think of a small dual 0-6, 0-6 transformer but all of this is highly speculative until some experimentation is done. One would have to feed on one side an AC test voltage e.g. 100mV RMS and trim/adjust the output for same reading on the gm meter. There remains of course the danger of transients produced from a faulty valve but without the DC current problem, a pair of back to back diodes might protect the meter on the other side of the transformer.
Hopefully, I will be clear of some outstanding tasks here in the next couple of weeks, then I can un-case my Sussex and try some more experiments. I have a gm meter calibration circuit I want to try anyway but this will have to be abandoned anyway if a DC isolating transformer is found to be required. The gm meter is already isolated from ground by the Paul Stenning transformer and this is a 100% requirement, so the meter damage must be caused by pushing the +/-200mV full scale over the top under fault conditions. It probably takes out the first op-amp on the meters pcb but being so fine a pitch SMD style, it is beyond my eyes and steadiness at my age to even think about trying a replacement. I did put the damaged meter somewhere, perhaps I should just get a new IC and ask one of the young lads in the wiring shop to replace for me, hopefully it will cost no more than a bottle of JD
__________________
Whether the Top Cap is Grid or Anode - touching it will give you a buzz either way! |
23rd Feb 2014, 7:50 pm | #1109 |
Tetrode
Join Date: Aug 2011
Location: Bristol, UK.
Posts: 67
|
Re: The "Sussex" Homebrew Valve Tester.
More good news for Sussex fans.
My tester has sprung into meaningful life once I removed the D16. I had a pair of new matched JJ EL34 to hand and found they gave Ia bang-on at 75mA and Gm just under 10. Previously it was fine on the other valves that I used - mainly audio stuff like ECC83 and 6L6. Could the variation in experience be explained by diode tolerance or mounting? My D16 mounted on top of the board on the pins next to the R12 resistor - It's just a thought. @Les, I don't think you need to hide - you have contribute such a lot to this project already. |
23rd Feb 2014, 9:10 pm | #1110 |
Hexode
Join Date: Dec 2009
Location: Leeds, West Yorkshire, UK.
Posts: 250
|
Re: The "Sussex" Homebrew Valve Tester.
Martin,
I can't readily think if I used different components between the three. The semiconductors were all bought at the same time so as to have several spares of each. Likewise with the meters, all from the same source. I'll have to have a look more closely at the three models to see if I can spot anything significant. Regards
__________________
Howard G7AJN/M3OCL "How hard can it be?" - Jeremy Clarkson |
23rd Feb 2014, 9:56 pm | #1111 |
Octode
Join Date: May 2009
Location: Linkoping, Sweden
Posts: 1,465
|
Re: The "Sussex" Homebrew Valve Tester.
I might be wrong here, but since we have a 10 Ohm measurement resistor which it will flow some 70-100mA through with an EL34 being tested the voltage across this resistor will be some 0.7V-1V which will be enough to forward bias the silicon diode. I think that this will affect the circuit in such a way that you will see these strange readings, at least when the current flowing through the measurement resistor is enough to forward bias the diode. The diode won't be affected by the AC-voltage from the amplified grid signal as long as it doesn't get the valve to oscillate, and with gm-values well below the required current.
So as Les says above two diodes in series might do the trick as that would mean some 130-140mA, or even three or four diodes - but then the DMM module won't be protected properly while high current valves are being measured. It would be possible to use circuits similar to the ones used in the AVO VCM163 for both the oscillator and meter amplifier with small transformers isolating the signals - or similar ones with modern IC's as amplifiers as long as the isolation transformers are being kept in the circuit.
__________________
Martin, Sweden |
24th Feb 2014, 8:50 am | #1112 |
Rest in Peace
Join Date: Mar 2011
Location: Basildon, Essex, UK.
Posts: 4,100
|
Re: The "Sussex" Homebrew Valve Tester.
I have a bit of a confession to make. I was going round with a capacitor testing to see if I could reduce the 50Hz pulses you can see on my traces and I found the HT smoothing could be improved by almost halving the ripple by the addition of a 47uf cap.
When I checked I had only fitted a 47uf in there originally, I cant remember exactly what happened now as I built that module over a year ago, but I probably intended it to be a temporary fit until I had ordered a 100uf. As this makes a significant difference to the "look" of the trace on the anode under high loads, what is the maximum size capacitor I could safely fit in this place with out damage to the HT winding and diodes. I was testing some EL84 valves last night and found that 4 out of the 5 I had all seemed to go into what looked like oscillation at or near the recommended grid bias. (It is very useful to have a scope on the anode whilst testing valves). Whilst the valve was in this condition the anode current and GM values were jumping up and down by a few mA. By giving the bias a bit more negative volts they tended to stabilise. I then found that when resting my arm on the patch panel wires that the oscillation stopped. Would this tend to indicate that the valves doing this are better or worse than the one which did not tend to oscillate. The anode current and GM values were not much different to the one that did not oscillate. I also tested some EF80's 86's and ECC83's and 81's and some of these came up with very good results. I also identified a couple where anode A was no where near the results for Anode B. (down to .3mA) With regard to EL34 valves and matching a set of 4 for use in say a 100 watt guitar amp, is it best to match the anode current or the GM value when pairing them up. Thanks Mike Last edited by crackle; 24th Feb 2014 at 9:08 am. |
24th Feb 2014, 5:58 pm | #1113 | |
Nonode
Join Date: Aug 2010
Location: Resolven, Wales; and Bristol, England
Posts: 2,608
|
Re: The "Sussex" Homebrew Valve Tester.
Quote:
I have tested some matched sets from Groove Tubes and AFAIK they do the same. Richard |
|
24th Feb 2014, 9:34 pm | #1114 |
Octode
Join Date: Mar 2006
Location: Watford, Hertfordshire, UK.
Posts: 1,270
|
Re: The "Sussex" Homebrew Valve Tester.
Hi Mike, these are the results I obtained a little while ago when testing a batch of 6CA7's. All four were very lively compared to an EL34 and this seems to be evident in use, one has to apply a bit more negative bias than usual. But the graphs show a pretty good test at high Anode current even though the diode is in place, maybe it has gone open circuit .
Looking at the graph though, something does not look right at maximum current and something I failed to see in the first place. I am wondering if the valves tested behaved themselves at the AVO test settings but then went astray with more current, so maybe the diode was playing foul after all? The four valves tested gave the attached results when tested with the AVO settings. These were taken as a spot measurement whereas the graphs were taken in steps of bias voltage over a period of time. I am going to look into the viability of fitting some kind of transformer in the Anode circuit and see how that goes.
__________________
Whether the Top Cap is Grid or Anode - touching it will give you a buzz either way! Last edited by Top Cap; 24th Feb 2014 at 9:52 pm. |
24th Feb 2014, 10:29 pm | #1115 |
Rest in Peace
Join Date: Mar 2011
Location: Basildon, Essex, UK.
Posts: 4,100
|
Re: The "Sussex" Homebrew Valve Tester.
I have attached the drawings I did for my Sussex Valve tester.
Feel free to use any for your own personal use. But do check the veroboard layouts against the schematic diagram as some changes were made during assembly. (I have done my best to updated the drawings but the previous statement is to help protect me in case I have made a mistake.) The main component board contained all 3 modules, I have separated them just for clarity. The relay board was assembled on a separate veroboard. Where high voltages were involved I removed alternate copper strips off. (Heated along copper strip with soldering iron then pulled up one corner and pulled) Mike Last edited by crackle; 24th Feb 2014 at 10:35 pm. |
25th Feb 2014, 12:36 am | #1116 |
Octode
Join Date: Mar 2006
Location: Watford, Hertfordshire, UK.
Posts: 1,270
|
Re: The "Sussex" Homebrew Valve Tester.
Mike, I just had a very quick look at one part of the circuit, that being the bias voltage. The diodes appear wrong and also the l
Leakage LED is not fed from the correct track. You probably have already spotted this. Les
__________________
Whether the Top Cap is Grid or Anode - touching it will give you a buzz either way! |
25th Feb 2014, 8:29 am | #1117 |
Rest in Peace
Join Date: Mar 2011
Location: Basildon, Essex, UK.
Posts: 4,100
|
Re: The "Sussex" Homebrew Valve Tester.
Hi Les
It is a good job I did not build it like that. I have corrected the drawing. see attached. Please could a moderator delete the Main veroboard layout in post 1115 above. Thanks Mike |
25th Feb 2014, 8:49 am | #1118 | |
Moderator
Join Date: Jan 2003
Location: Ipswich, Suffolk, IP4, UK.
Posts: 21,288
|
Re: The "Sussex" Homebrew Valve Tester.
Quote:
__________________
Graham. Forum Moderator Reach for your meter before you reach for your soldering iron. |
|
3rd Mar 2014, 9:44 pm | #1119 |
Rest in Peace
Join Date: Mar 2011
Location: Basildon, Essex, UK.
Posts: 4,100
|
Re: The "Sussex" Homebrew Valve Tester.
Hi
I thought I would just update you with developments with my valve tester. On Friday I took the tester round to Westcliff's house to do some comparisons with his AVO valve tester. The comparisons showed that my tester was down on the GM readings on all the valves we tried. I have been aware of the fact that each time I try to check the oscillator output voltage it seems to measure something different. I had a plan to add a change over switch to switch the GM meter from its GM function to measure the oscillator voltage and move the preset control to the front panel so it could be calibrated on each occasion I wanted to test valves. The 200mV AC meter used for the GM measurement seems to measure the oscillator signal more reliably than my DVM and oscilloscope. When I checked the oscillator voltage with the GM meter it was only about 85mV so I reset the oscillator voltage using the old GM meter. However I found that the preset pot was far too sensitive so I fitted a 10 turn 10k pot. Not sure if it was necessary but I also changed R21 from 470R to 1k. At this point I decided against fitting the pot and switch on the front panel, but instead to change the pot into 2 fixed resistors. (this way I cant keep fiddling with it.) I measured the resistance from the wiper to each end of the track and the results were 8.81k and 1.26k so I used a 7.5k and a 1.1k resistor to form the new potential divider. The results were, 101mV slowly rising to 104mV after 15 min or so, close enough I think. I have found that on my DMM the AC milli voltage is influence by the bias voltage and this is possibly why it was only 85mV when I started these tests. I have since tested the same valves again and the GM readings are now very close to those obtained on the Westcliff’s AVO tester. When testing EL84 valves quite a few of them, when fully hot, seem to go into a state where the current is jumping from 41mA to 56mA and the GM is jumping from 10 to 3 or 4. This is cured by holding the anode patch cable in my hand and it returns to a steady reading.. Would ferrite beads help with this situation. I also tested some rectifier valves today and these tests also went well. Not sure if I mentioned this before but I have fitted a means to disconnect the AC signal from the grid and inject an alternative signal through the patch panel. So to sum up, it all seems to be working well and valves which are expected to be close to the manufacturers specs are reading as expected. Mike |
4th Mar 2014, 3:23 pm | #1120 |
Tetrode
Join Date: Aug 2011
Location: Bristol, UK.
Posts: 67
|
Re: The "Sussex" Homebrew Valve Tester.
Thanks Crackle for the update. I recognise that drifting oscillator too when I had D16 in place. I haven't checked after removing but I will. As I suspect that is why my GM were all about 9.5-10 when I expected more like 11+
I saw your note about the alternative input source: Not sure if I mentioned this before but I have fitted a means to disconnect the AC signal from the grid and inject an alternative signal through the patch panel. Was that just a part of debugging option or do you have plans to be able to do listening tests or insert an separate oscillator. It sounds intriguing either way. |