Klausmobile Tube Tester Files
What a tube tester should be - An Afterthought
After publishing curve tracer schematics I received a few emails asking question, if I would sell PCBs, kits or otherwise commercialize this design. While it is definitely not my intention, I also realize that this design has certain flaws that a commercial curve tracer must avoid.
The main flaw is, it is a SMALL SIGNAL TUBE curve tracer and a MEDIUM POWER TUBE tester/matcher. Although I cannot evaluate even EL84's linearity at it's specified DC point, I can say if the tube is working or not, and match a pair at some lower U/I point. If the two tubes have identical bias, gain and transconductance at half current in triode mode - they must match at normal current just as good, triode or pentode wiring. Same applies to 6L6, 6AS7 and some other tubes in 10-25W plate dissipation class. But that's as far as it goes, it won't test bigger bottles like 2A3.
A. An all-round commercial tube tester must test POWER TRIODES. Obvious, isn't it? Plate dissipation power is the easiest thing to upgrade - simply replace the plate supply for something beefier. But that's where easy solutions end. Because power triodes need grid bias sometimes at -30, sometimes at -100V. Testing them at 0-10V bias is quite meaningless. Grid Bias Must Go Lower than an opamp driver can supply (-10V). Which means - a separate -150V supply and a bias amplifier (regulator). This isn't as easy but can be done without changing the A/D interface.
B. An all-round commercial tube tester must test PENTODES as PENTODES. Which means a third remote-controlled power source. Technically, it can be implemented as just another plate requlator, powered from the same rectifier. Oops - low-cost A/D boards have only two DACs, which are already used for plate/grid. Costs of a separate DA board makes the whole idea prohibitively expensive for a DIY market. The solution: forget off-the-shelf D/A boards, make a custom DAC/ADC interface (with digital I/O ports, controller etc) and hook it up to PC through a standard serial cable. Easy? Well, not my cup of tea, honestly.
C. An all-round tester must employ very high precision ADC - that's simple, instead of 1:50 (mA) current range, we're talking 1:500, and to make lower end measurements reliable, '1mA' reading must be at least 7 bit precise (10 uA LSB resolution). This requires full 16 bit SE conversion. Such precision is impossible in PC-based boards unless they are fully opto-isolated. Another point for relocating A/D conversion into the tester cabinet, away from the noisy PC innards.
OK, 500mA (6C33C) is too high for a practical design, let's limit current to 200mA @ 300V. In the end, we have a black box including:
Very close to Audiomatica Sofia specifications, isn't it? These Italians, they knew it from the very start :))
English Home - Russian Home - English Tube List - Russian Tube List - Mail (c) klausmobile 2001
