This is a description of building a tube tester. It's nearly finished by now but maybe some of you guys more knowledgable with tubes have some thoughts. This is as much for me to order my thoughts and write down how the thing works as it is for anybody who's interested.
It's going to be quite wordy
.
I don't do much with tubes, only the occasional small audio project or the odd tube radio for repair. A few weeks ago a friend asked me to repair his old tube radio which stopped working. Nothing special, 70 year old caps that needed replacing and a couple weak tubes. It's alive again now even with a bluetooth receiver. But that got me thinking again of a project I had in mind for quite some time. I "need" a proper tube tester! I did not like testing the tubes in that radio using 4 power supplies and loads of test leads and meters. It's so messy.
I built a very simple tester years ago, but that thing is so fugly and huge I'd be ashamed to show it. Also it's just not very good and has been sitting in the shed for 20 years.
Buying an old vintage tester was not an option. These things are either hugely expensive or do not meet my needs and are still expensive. I did not want a simple go/no-go emissions tester. I wanted a proper one, like the Neuberger RPM375 or similar vintage machines but more comfortable. I had no intention building the supplies with huge rheostats or such nonesense.
So I started thinking about it and the first challenge - and one of the reasons the first one was so huge - was the power supplies and more specifically the transformer. After a long time looking for commercial trannies that met most of my requirements I decided to wind my own. Something I really don't like to do but there was no other solution. Other than using multiple separate transformers.
I wanted to avoid that, as I decided early on the tester had to be as compact as possible, after all I don't use it that often and so do not want to waste a huge amount of space. First thought was to use a 3HE 19" rack as enclosure, but that turned out to be impractical. In the end I decided to use an old aluminum "suitcase" I had lying around wasting space. It seems there was a reason most vintage testers came in that form factor.
Next step was to decide what technical data I wanted to achieve:
-Anode voltage: 0...400V DC 150-200mA, regulated with current limiting
-negative Grid: 0...50V DC, unregulatd
-Grid 2: 0...400V DC, 50mA, regulated with current limiting
-Filament voltage: 0...25V DC 2,5A regulated
Meters for anode voltage & current, Grid 1 voltage & current, Grid 2 voltage & current, Filament voltage & current.
In my stash of old transformers I found one with a SM102a core that seemed to be a good solution. The SM core is easy to take apart and rewind. A few back of a fag packet calculations were promising, the core should be able to handle enough power for my project. So I took it apart and unwound the old secondaries. I've kept the original primary.
I've built a small manual winding "machine", just a few boards screwed together with a threaded rod with a crank and a wooden spacer to hold the core bobbin. Past experience taught me that winding a transformer by hand is messy and the results are less than ideal. I've added a micro switch to hook up to a counter. That made keeping track of the amount of turns relatively easy.
As I do nat have a way to soak the tranny in laquer in a vacuum, I did the next best thing and applied liberal amounts of laquer with a brush while winding in an attempt to keep the hum down. It worked ok. I used kapton tape as interwinding isolation.
After every secondary winding I reassembled the transformer and tested them under load. The transformer now has six secondary windings.
One 320V for anode voltage, 300V for grid 2, 38V for grid 1, 27V for filament and auxiliary (relays) power, two 15-0-15V windings for auxiliary power in the power supplies.
While all this was going on I researched how vintage German tube testers worked, looking at countless schematics and user manuals. It took me three weeks to come up with a plan and first schematics. At first I wanted to use digital panelmeters but that turned out to be a huge pain in the behind so I went with analog meters - so much easier. Not that analog meter movements a trouble free, mind you. Previous experience with cheapy chineseium meter movements was bad. Most class 2.5 meters are just utter shite. Not that I need the accuracy as such, but still I decided to go with good quality vintage meters. I've bought a whole box of old meters on ebay years ago, nice Gossen, Neuerger and Simpson meters and some good quality USSR stuff. For good looks I did not want to mix different brands, so decided on the Gossen meters. Of course I did not have enough but this particular style shows up on ebay regularly here as Rohde & Schwarz, AEG and who know who else used them in their gear.
Also at first I planned to use a matrix of rotary switches to select the tube connections but it turned out to use too much space on the front panel. For a short time I thought about building a plug board, like Neuberger or Funke used them. I would have had to build that from scratch as it it nearly impossible to buy something like that today. Ghilmetti still has some in their catalogue, as it turns out only on custom order and hugely expensive. Relays were out too, too much trouble and how do you set them? Using a µC. But then I'm half way at building computer controlled tube tester, not what I wanted.
I could build this then:
http://www.roehrentest.de/So in the end I decided to use 2mm Banana sockets and plug leads. Not very comfortable but very space efficient, flexible, cheap and easy.
Another problem with rotary switches is the current & voltage ratings on the cheapies (Lorlin) is not all that great. Not sure how long they'd last with 400V and 200mA on them.
Another hard nut to crack is testing for shorts on the tubes before the proper test. The commercial devices of the past apparently used special switches to accomplish that.
This homepage provided very usful insights:
http://www.hts-homepage.de/RPGSelbstbau/RPG.htmlI've decided to use the version with SPDT switches the author describes. I used relays instead of single switches.
I think it works but have so far only been able to test one tube that has a broken filament and a short between filament & cathode. Not sure yet how good my implementation really is. It did not find the short on that tube because it is on the wrong "side" of the broken filament. Clearly more thinking needed, do commercial testers find all faults? I'm not sure they do.
I then started designing the basic power supplies. The board underneath the transformer generates 24VDC for my relays and the adjustable filament voltage with presets for 6,3 & 12,6V using a simple LM350T. It's very close to its maximum power rating, so to help it out I added a resistor in line before the LM350 to drop some voltage & dissipate some heat. Not very efficient but so far seems a good enough solution here.
Also on this board is the rectifier, caps & sundry stuff for the negative grid voltage. This is unregulated, just using a pot as voltage divider. I deem that to be okay as the grid takes virtually no current.
On this board are also the rectifiers & main caps for the two high voltage supply Va & Vg2. At last there's that huge relay at the bottom of the board. This switches 300-0-300 VAC to the test circuit instead of DC. I hope this will work for testing rectifier and diode tubes. In my mind it does, haven't been able to try: I have no glass rectifiers in my stash.
All the voltages are adustable by 10 turn pots, simply because I have dozens of the things and I like them.
In the middle of the unit is the big anode current meter with range switching and underneath the main rotary test switch. This switch lets you select testing for shorts, checking the filament, the main test and test for diodes and rectifiers. Instead of using a specially made switch or one with many wafers I decided to use relays. 14 of them to be exact. The switch sets the relays using a diode matrix.
Next to the main meter are the test lamp and buttons for vacuum tests.
Then on the right there are the test sockets. I decided to keep to the most common sockets because space and I do not see myself collecting ancient tubes. Underneath are fuses for the secondaries and the plug sockets that wire up the tube.
Then filament voltage with associated switches and meter.
The left side has pots, switches and meters for anode voltage, -Vg1 and Vg2. Also outputs for the various voltages. The two BNC's are for feeding a signal in and out to check a tube for microphonics. I'm not quite sure how I'll approach that yet.
Also there's an external input for grid 3. I decided against a integrated supply for that, as I've never needed it. And also: space.
The two boards underneath the left side contain fully regulated and current limited supplies for Va and Vg2.
Things still to do: Hook up the load resistors for testing recifiers, buying a rectifier tube,
hooking up Va current limiting properly, setting current limiting properly. 
Home
Help
Search
Login
Register
