Radio Restoration

09/01/2026

My HP400E AC voltmeter troubleshooting and repair article is now posted to the SPARC radio museum website, here:https://sparcradio.ca/wp-content/uploads/2026/09/Troubleshooting-and-Repair-of-an-HP400E-AC-Voltmeter-with-Appendix.pdf .

08/31/2026

During the troubleshooting of the HP400E AC voltmeter, I used an ATLAS peak DCA55 semiconductor tester. I have had this unit for well over a decade and it has provided stalwart service.
I have just watched a YouTube video of someone repairing one of these units and was very impressed by the customer service he received when he asked if they could provide a spare part for him. To view this video, see here: https://www.youtube.com/watch?v=4dxwOpe0vE0

08/31/2026

I've had a rather busy weekend, so not a lot of radio-related stuff happening (apart from checking out the clone Tektronix 'scope with my friend) - he is now giving the chassis away for free (recipient must collect!)...
I did managed to spend a couple of evenings putting together an article on the HP400E AC voltmeter troubleshooting and repair - some sample pages from the draft article are attached to this post. I will post a link to the article once the article has been uploaded to the SPARC radio museum website.
I am hoping to find some time to carry out some further checks on the Hickok 539C tube tester this week...

08/30/2026

I had a call from my friend who has the Tektronix 545 clone - he said that he had removed the B+ relay (T601) from the low voltage power supply and, as I had suspected, it appeared the break in the the solenoid winding was to one of the solder joints on the base of the solenoid. Thinking this might be repairable, I called by his house this afternoon on my way to pick up Wyatt for a walk in the park.
Sure enough, one of the connections between the solenoid wire and the solder lugs was missing the wire - it also looked like someone had tried to repair it as the soldering looked very poor compared with the other (good) joint.
I removed the tape from around the solenoid bobbin and managed to tease out the missing end of the (very fine) solenoid wire - I would say 40SWG. With some care, I stripped the enamel from the wire and re-soldered the wire to the solder lug. However, the coil resistance measured only 20ohms, not the 2Kohms specified in the parts list. A 20ohm resistance for the size of the bobbin of 40SWG wire forming the solenoid seems to be unrealistic, and I tried passing up to 120mA through it, but there was no hint of the armature moving.
In the 'scope circuit, the solenoid is operated from a 180vDC supply via a 6Kohm series resistor. If the solenoid had the specified 2Kohm resistance, this would result in an operating voltage on the relay of around 45v and a current draw of around 23mA. If the solenoid only has a 20ohm resistance (as I measured), the current draw through it and the series resistor would be a little higher (30mA), but the voltage drop across the solenoid would be negligible (0.6v), and this would certainly not operate the relay. My conclusion was that the relay solenoid has likely developed an internal short.
The owner had decided to give up on the 'scope anyway, and has started to dismantle it for parts - so far removing all the tubes (around 80 in total, some having Tektronix labels - strange to find these is a clone!), and some large power resistors. I recommended that he also keep the power transformer and several of the precision resistors (strangely, made in Canada), and some of the connectors and switches.
We spent some time powering-up some of his real Tektronix 'scopes: a 535, 545 and a 503. All powered-up ok, though as he had noted the CRT in the 503 is rather dim (pity, as it has a nice blue phosphor).

08/28/2026

I spent some time at my friends house this morning checking out the clone of a Tektronix 545 'scope, now with schematic and manual in hand.
Following several voltage and resistance checks on the low voltage power supply circuits, the likely culprit seemed to be a faulty relay (K601) that switches several B+ supplies on after the tubes in the 'scope have heated-up.
The solenoid of relay K601 is operated after a time delay of 45 seconds, effected by a thermal time delay relay (K600, Type 6N045T) switching the K601 solenoid voltage.
The thermal time delay relay (K600) checked out ok, as did the circuitry between the solenoid of K601 and the +180v unregulated B+ line supplying the +100v regulator circuit. This only left the solenoid of K601 as suspect, and sure enough it was open-circuit. The parts list specifies a 2000ohm DC relay coil for K601, and the relay has a 4PDT contact set.
To see if this was the only issue, I operated the relay manually by pushing the K601 relay solenoid armature until the contacts switched. Doing this caused a brief flash to appear on the CRT, but also a noticeable change in the running speed of the cooling fan(!), indicating a severe overload on one (or more) of the power supply circuits....
So, it looks like there will be more problems to resolve even if a replacement relay can be found, A possible cause is a shorted electrolytic or paper dielectric B+ filter capacitor, but that troubleshooting is for another day (if at all....), as I ran out of time.

08/26/2026

Out of interest, I checked the bias voltage settings on my STARK 9-66 and Hickok I-177 tube testers. These units does not have a meter to read out the bias voltage, but simply specifies a division number on the bias control.
Attached to this post is a table of the bias voltages for each five division increment on the STARK 9-66s bias control (0-100), along with representative 'scope traces of the bias voltage - this shows an unsmoothed full-wave rectified DC bias, though at low settings the waveform is more like a sine wave, and a table of bias voltages for the Hickok I-177 bias control (0-80).

08/26/2026

As I was walking Wyatt along the coast yesterday afternoon, I received a phone call from a friend about a vintage 'scope he was working on, and asked if I could call by take a look. Wyatt did not seem too perturbed by this, so I agreed, and headed home via my friends house (I think Wyatt quite likes going there as he has a dog-friendly enclosed yard he can explore).
He was working on a 1950's 'scope that is a clone of a Tektronix 545. This 'scope was manufactured by Lavoie Laboratories Inc. (Morganville, N.J.). Other US manufacturers also produced these clones, including Hickok.
These cloned Tektronix 'scope represents an interesting piece of test equipment history, as it was the subject of a court battle over patents that started in 1961 and was not fully resolved until 1978(!). I understand that the cloned 'scopes were ordered to be destroyed, but I guess a few managed to escape...
The Vintage Tek website includes the full story is anyone wants to check it out (see: https://vintagetek.org/clone-scopes/ ).
The fault in the clone my friend owns appears to be in the high voltage supply as the five rectifier tube heaters are not lighting up, and their heaters are each derived from a half-turn winding on the high voltage transformer operating at 60KHz from a 6AU6 oscillator. I did not have much time (or a schematic to work from), so after deducing this, I headed home and downloaded the manual for my friend. He last used the 'scope over a decade ago (when it was working well), but is now looking to sell the 'scope as part of a collection downsizing effort, so hopefully he can fix the problem (if not, I may be spending some more time in his basement!).

08/26/2026

As a small postscript to my last post regarding comparing the Gm results obtained from two tube testers for the same tube, and comments on 'finicky' bias adjustment controls, I came across a YouTube video where a Hickok 539B is compared to a Hickok 539C, and the bias control on the 539B shows a good example of such 'finckyness'(!). The video link is here: https://www.youtube.com/watch?v=2iCxIHkLgI0 .
One of the comments to that video is also interesting and supports my musings regarding the setting of the bias and AC line voltage controls: "...If you want accuracy and repeatability, the AC line adjustment is most critical, if it is offset as much as 1 division the GM can vary as much as several 100 umho’s. The bias voltage is also very critical, a few tenth’s change will affect GM. It is very important after pressing P4, that you readjust your line and then the bias to the correct value before taking your GM reading . (Both will change when P4 is pressed and back when released. (You can’t just set and forget.)...".

If you want to skip the intro, start somewhere around 2:05Results...

08/26/2026

I have been rather busy over the weekend, apart from a little time on Sunday morning, so no 'quality' time for any radio related stuff. I have been asking around to see if anyone locally owns a Hickok 539C tube tester to compare measurements made on a number of tubes with directly, but unfortunately no success. However, one of my friends who lives close by has several other testers though, including a TV7 and a Marconi MU-101.
The Marconi MU-101 is a solid-state version of the STARK 9-66, refurbished by Marconi (I think in the 1970's). I called by his place on Sunday and tested some of the same tubes I had been testing previously (per previous posts) in in the MU-101. The results are included in a table attached to this post, along with comparisons between EF86s, 6V6s, an EL84 and a KT66 on the two mutual conduction testers I own, ie. a STARK 9-66 and a Hickok I-177 (I chose audio tubes as I understand that is what the owner of the Hickok I-177 will be using it for). The I-177 does not seem to have tube data for some types, e.g. EF86, so there are no results in the table for those tubes for that tester.
The attached table includes percentages (Gm results compared to the STARK 9-66 average new tube value) to allow easier comparisons. It’s interesting to note that for the dual triodes, the Feb 1957 issue of the RCA Electron Tube Handbook quotes two values ('series' and 'parallel'), and the same one is not always quoted in the STARK roll chart. I have also included the Hickok minimum acceptable Gm values from its roll chart, and an average new Gm value back-calculated from this using the 65% of average new value for the minimum acceptable Gm value I understand Hickok used. There is significant variation between the average new Gm values from these various sources (maybe the RCA value should be the 'definitive' one?).
I have read that a +/-10% variation in Gm value between mutual conductance testers when testing the same tube would be reasonable, though I have not seen any data to back that up. When other things are taken into account, e.g. age of the tester components, accuracy of the actual meter movement (eg. due to degradation of the magnet), slight variations in control settings/k**b adjustment errors, etc, I think even +/-20% would be optimistic, everything else being equal. However, this does not account for the actual test conditions encountered by the tubes in the different testers(!) – see below...
My testing efforts showed that by far the greatest variation in Gm values obtained for any tester I used was the result of the bias voltage applied to the tube, and I found that the bias controls are rather 'finicky' (likely due to wear in the pot), especially at the lower end of the range, where just touching the bias setting k**b slightly can change the Gm value reading obtained significantly.
However, maybe more significant than these factors is what exactly the 'bias voltage' is. In an amplifier circuit, this is, of course (or should be) pure DC, with the (AC) signal superimposed on it. I checked the bias voltage on the three testers I have here (the Hickok 539C, the STARK 9-66 and the I-177), and I was really surprised at what I measured using different voltmeter types and the waveform displayed on a 'scope. I have attached some photos of the waveforms to this post as examples.
My conclusion from these observations is that I should take any Gm value measurements made on different testers 'with a pinch of salt' (almost), as how the testers apply the bias, and the 'signal' (as in the case of the Hickok 539C) is rather arbitrary, and bears virtually no relationship to a real world application. Also, most DC meters do not measure the true DC value of a complex half-wave rectified waveform, and most AC meters do not measure the true RMS value of an AC waveform unless it is a sine wave.
I tried measuring the bias voltages (DC and AC) using an HP3456A, Bryman BM786, and a digital 'scope (Siglent SDS 1202X-E), and the AC component only on an HP3400A (which does measure the true RMS value of complex AC waveforms). I have attached a comparison table to this post that shows the voltage readings of the different instruments used at various bias voltage settings as read on the Hickok 539C – they are ‘all over the map’, which is not surprising as there is a complex AC waveform superimposed on unsmoothed rectified DC.
The bias voltage waveform observed on the 0-10v range (a very complex waveform) of the Hickok 539C is completely different from that on the 0-50v range, and probably warrants investigation. Given this, I can totally understand why tubes that require a low voltage bias setting on the Hickok 539C, such as the 12AX7 at 1.3v, may have issues in obtaining a reliable Gm value (as the owner was finding). It is also interesting to note that RCA specify a 1v bias for the stated average new 12AX7 tube Gm values, not 1.3v per the Hickok 539C tube roll chart. A 1v bias setting gives a much more reasonable Gm value for this tube type, eg. for the Marconi 12AX7 tube referenced in the table, the Gm with 1v bias is 1200/1375, ie. very close to the RCA spec., as compared to 875/900 at 1.3v bias.
In summary, I am amazed that the various commercial standard testers give as reasonable results as they do, as they are certainly not testing tubes under ‘real world’ conditions, or even in the same way. The bias voltage applied varies from almost a sine wave to a complex AC waveform superimposed on an unsmoothed DC voltage. These variations, along with age, adjustment and operator considerations noted above really brings into question the validity of any comparison of tube measurements between testers. In future, I will have much less confidence in the Gm results obtained, especially for small signal tubes, and will be using testers mainly as a 'failed/weak/strong' indicator. An interesting article on comparing a 12AU7 tube measurements on a large number of tube testers can be found here:https://pdf.dzsc.com/88888/200848104729603.pdf
I plan on carrying out some further checks on the Hickok 539C over the next few days as time allows (I have a busy week ahead), probably focusing on the bias circuit, but also checking voltages and carrying out some of the calibration checks described in the Higgins-Roper tome: https://www.byan-roper.org/m_higgins/2017-539bc-Calibration.html .

08/21/2026

I changed out the two HP400E AC voltmeters this morning, so the newer unit is back in its rightful place in my instrumentation stack. I ran a few more tests on it after that and the calibration (amazingly) does not seem to have been affected by changing out all those capacitors, resistor R21 and all the semiconductor swaps (though the originals are now back in place).
I took the lid off the older unit once I had removed it from the stack and R21 in that looks to be a much higher quality part (porcelain body) - I would say at least a 1W part, so should be ok. I decided to leave 'well alone' for now, and packed the unit back into a storage cupboard.
As a footnote, there is a note on the HP400E schematic that states "R21 is emitter load for Q5 and Q7" (obviously not strictly true, as Q5 is a JFET - so is the source load for that device). Yes, even HP didn't always get things correct...