Tweet
There are many models of transistor/IC testers which measure breakdown voltage, DC HFE/gm curves, saturation voltage, etc. What about tube testers? Did anyone build a "production" tube tester that generated the gain curves (plate and/or transfer characteristics) for a tube? How do current tube suppliers "match" tubes?
-
-
soldering stuff that's broken, breaking stuff that works, Yeah! -
Get out yer plastic.....The farmer takes a wife, the barber takes a pole....Comment
-
Engineers need curves, production and maintenance/repair people do not. Tube testers are for repair guys. They made curve tracers back in the day, or a circuit was made up for scope display or to laboriously generate data for graphing. But not generally in a "tube tester."
Tube sellers basically set up a tube tester type circuit. They apply a particular set of voltages to the tubes and loads they determined their own ways, then they stick each tube into the setup and measure whatever they want, usually gm and cathode current. The conditions don;t necessarily match your amplifier, but they are consistent in the test lab. They sort the tube so that ones with the same numbers get grouped into "sets." They are not matching curves, they are matching just a couple crude bits of data.Education is what you're left with after you have forgotten what you have learned.Comment
-
Most tube manuals will give you curves for typical examples of tubes but if it's a curve for the one you have, then the field of ready made devices is pretty limited. Alan Douglas' book Tube Testers and Classic Test Equipment discusses tube curve tracers (Tek, Fairchild) and also shows some ways to use a scope to generate curves. There is one tester currently on the market that can do curves, can't remember the name offhand but it's spendy-about $2,500 I think. There's a series on Instructables about building your own tube curve tracer. Of course you may have needs for these things when you consider that tubes are non linear.
Tube Curve Tracer Ver 1.1
It would probably be a fun thing to build this winter.
Your average good quality tube tester applies a modest voltage to the plate of whatever tube is tested, but as many have observed, that doesn't really put them under the kind of stress that you'll find in a modern amplifier. My Hickok maybe makes a couple hundred plate volts on a good day, when a 6550 in an SVT is hitting 660v on the plates. So in addition to determining mutual conductance a second set of tests is called for, and that is evaluating plate current under something approaching working voltages and appropriate bias voltages. That's where the real matching comes into play, because plate current is far more important than gM-as long as you've got something above nominal for a new tube you're fine there.
I'm a believer in carefully matched, selected and burned in tubes for amps that run high-ish internal voltages and where reliability is critical.
More on this subject. People have taken garden variety emission tube testers and turned them into curve tracers. This looks promising.
http://www.dutchforce.com/~eforum/in...howtopic=28929
I share km6xz's cynicism concerning tube rebranders and current day makers.Last edited by Prairie Dawg; 10-20-2011, 05:05 PM.Comment
-
Curve tracers are available for low cost surplus that fully characterize semiconductors. Some of the most available at a reasonable price is the Tektronix 575-577 series. You can find them in good condition for less than $500, a fraction of their original value. Some have the option of peak collector currents of over 200 amps.
Tube curve tracers are more needed now than ever since the curves in published specs are bogus and are copied from tube manuals from 40 years or more ago and have little to do with what is being sold today under the same tube designation. Tektronix made a classic all tube curve tracer called the 570 that is pretty expensive and rare. The tube re-branders do not bother with doing their own curves since they either do not care or don't know how...I suspect the latter.
I have two, one is useful for a range of devices including tubes and transistors, the Fairchild 6200b which can sweep collectors/anodes to 1000 volts. To generate curves for tubes, only a heater supply and screen supply are needed. I find them extremely useful in repair and design to be able to discover exactly what the part I am dealing with actually is.
There are cheap, lower precision transistor curve tracers available used... Heathkit and Eico both made low cost but effective curve tracer adapters that used your own oscilloscope as the display unit. They go for around $25-50 at ham radio swap meets. For a hobbyist, they are all that is needed. Heath made some very cost effective test instruments back in the heyday of hobby electronics in the 50s and 60s. Their VTVMs, IG5218 audio generators, and many other items are still sought by bench techs. In fact, their old IG5218 signal generator is the most effective all in-one box I have used for generating, metering and frequency selection of audio test signals from any company. Although I have a bunch of top lab type gear, I buy those generators for work benches every time I see one for sale.Comment
-
That sounds like a pretty handy ass little tool there Km6xz and wish I had one here in the states to troubleshoot with. Something else about tube testers that some of the younger guys may not realize but everything or almost everything that we can do now in SS was done back in the day with tubes so most of the testers were for every kind of tube almost out there. Many different voltages were used which is why those tube tester transformers are so cool because you can almost dial in any voltage you want for testing using different taps. Most tube testers were all about grid shorts,heater shorts and if the tube was conducting. Many really didn't use gm or mu or even how long the tube was going to last it just checked to see if it was operable.KBComment
-
since i've collected a few thousand tubes in the last 35 years, it made sense for me to buy a few tube testers. i love my maxi matcher for output tubes (with the EL84 and 7027A adapters). it's great to have a quick way to match and find the duds.
for preamp tubes,and everything else i have an old triplett 3444 that's always worked well, but to check preamp tubes for microphonics i just plug them into either my showman (12AX7's/12AT7's), my SVT or V4B (12DW7's/12AU7's) or my B15N (6SL7).Comment
-
Originally posted by km6xz View Post
I like the idea of adapting a 576 to generate tube transfer curves - the Tek 370a would be an even better choice but more cost ($400 vs $2K). When the photovoltaic industry was shining it seemed that curve tracers were in demand again but now there are dedicated instruments for module and array characterization.
It's pretty sad if tube suppliers are not characterizing product and then matching them by using a single bias point - like defining a line with one set of coordinates.Comment
-
It's pretty sad if tube suppliers are not characterizing product and then matching them by using a single bias point - like defining a line with one set of coordinates.
Sometimes if you're in the field that's all you get . There's one guy who's converted a garden variety emissions tube tester into a curve generator with output to a scope. It looks like something I may try and build this winter. I think it could yield some seriously useful information and could be a helluva lot of fun as well.
If you've got a pretty good idea of what your plate voltage and bias voltage is supposed to look like you can come up with a reasonably good "one size fits all" matching. It's really a good thing that tubes are so forgiving.Comment
-
Yes, a good precision curve tracer that sweeps anode/collector/drain-source over a wide range, plus positive and negative current and voltage steps for grid/bases/gates and display 10 trace curve families is a handy tool. It is hard to have marketing hype overcome objective data if it is available.
I have been using the Fairchild for generating plots for my own brand of tubes we are trying to get going here in Russia. I include the actual reproducible curve family plot for each tube by serial number.
The old tube testers had heater transformers with many taps for 1 volt on up to 115volt and everything in between. Screen voltage on the mutual conduction testors was usually set with a rheostat and a couple used variacs for anode voltage. My old Lambda regulated tube power supply uses variacs for screen and plate voltages, and a rheostat for the bias supply and only 5volt, 6.3 or 12.6 volts AC options for heater. I used that for manually plotting curves or testing breadboarded circuits out before getting the 6200b. My next project is building a computer based curve tracer for directly printing the curves and printing the serial number sticker automatically.
The biggest problems with units like the 570 or 6200 are weight and size....besides cost. I shipped my 6200 from my home in California, for $266 by the cheapest way, USPO. UPS wanted just under $1000. All my other test gear here in Russia was brought to Russia 1 or 2 at a time in my suitcases on each return trip back from the US.Comment
Comment