I'm interested in being able to do the following: (1) Choose a vacuum tube at random from the RCA Receiving Tube Manual. (2) Based on its characteristics, figure out a semiconductor device with the same characteristics. This may also require designing a circuit containing the semiconductor device.
How does one do this?
There are probably tables that tell you what semiconductor device to use in place of a given tube. I tried to search for semiconductor replacement for vacuum tubes and only got hits about the history of semiconductor devices, which as a whole replaced vacuum tubes, but no tables of specific replacements.
Also, is there a book comparable to the RCA Receiving Tube Manual for the relevant class of semiconductor devices?
Allan Adler has been on usenet for years. He seems rather similar to "Radium". He asks daft, naive questions, and doesn't seem to do much with any helpful responses. I stopped taking him seriously ages ago.
Idunno about "ages" ago, but I've begun giving him the same treatment.
I mean, come on... He claims to be a complete electronics newbie, yet he has grand plans to build semiconductor equivalents of tubes? Nothing like trying to do calculus before learning how to add and subtract...
Don Bruder - firstname.lastname@example.org - If your "From:" address isn't on my whitelist,
or the subject of the message doesn't contain the exact text "PopperAndShadow"
At least one difference, however, is that his post was well stated and realistic. In contrast, Radium is a blatant troll.
I have several Boatanchors for which I will be building solid-state replacements for VTs this winter, so I am sympathetic to his wishes. So far their designs exist only roughly in my head, but they are very feasible.
It appears that when some people don't know the answer to a reasonable question, they instead devote their efforts to preventing it from being asked, especially when the question comes from someone who asks a lot of such questions.
Let me try to get this straight: do we or do we not like talking about electronics on this group? And is it or is it not entirely voluntary on the part of readers whether they choose to answer a question? That being the case, I don't see what the problem is.
Some people who don't understand why I might ask a particular question assume that a legitimate question must be part of an ongoing, hands-on project, with real parts and equipment. I'm not aware of that being one of the official premises for this newsgroup. Others, lacking any actual information, invent scenarios which make sense to them based on the amount of information the actually have, which might be very little; I don't always explain all of my reasons for asking a question. Then they get mad because their invented information offends them.
In the present case, my question has the following origins; I didn't provide this background since it was irrelevant to the answer to my perfectly general question: (0) Motivated by Michael's advice that I do a lot of reading, I started looking through some of the literature I already have in my possession but have been neglecting. (1) I happen to own an old Beckman DU-2 Spectrophotometer (the model has been changed, with the same model number, since this one was made). It has no power supply. I have a lot of documentation for the old DU-2, which includes circuit diagrams for the power supply. Every once in a while, I take the docs out and wonder whether there is any way to get the thing operating. On this particular occasion, I started looking at the tubes, some of which are listed in the RCA manual, and some are not, being specially made for Beckman, and therefore undocumented. (2) Then I started wondering whether one could modify the schematic so that the tubes were replaced by circuits containing semiconductor devices and, if so, how one would do it. (3) Then I realized that, although I have seen it remarked generally that semiconductors replaced tubes, I wasn't sure whether each tube had its semiconductor replacement. One reason to think it might is that I do vaguely remember seeing tables of replacements for particular tubes in old books, maybe old Radio Amateur Handbooks, but I'm not sure. (4) At any rate, this raised the general question of how one would figure out how to replace a given tube with a semiconductor circuit. And since I don't know how one does that, I asked.
This background will perhaps clarify the reasons for my asking the question. The people who don't know the answer still won't know the answer and the people given to road rage on the information superhighway will do what comes naturally to them.
When I was in Wolfenbuettel, I saw a very old house on which was written (some spelling errors are mine, others in the original): "Geh dein Weg und lasse the Leute reden."
You're going about this the wrong way. Instead of using the original circuit, and trying to find semiconductor replacements/equivalents to the valves given, (which would incur such a radical re-design that you'd be as well to start from scratch anyway) look at the requirements of your spectrophotometer in terms of power supply: what voltages/currents are needed in order to make it work? Then design yourself a nice linear supply with modern regulators appropriately heatsinked, and allow a generous 50% current output overhead. Simple, effective, minimal components required, and it'll get your spectrophotometer up and running with far less brain strain than sticking to the original PSU circuit.
I agree with you completely and had reached the same conclusion before posting my question about replacements for tubes. However, I thought that the question about replacements for tubes was interesting in its own right and decided to ask about it.
To describe the origin of my question is to describe a historical process. It is not the same as describing "the real question". One of the reasons I omitted that history from my original posting was precisely to avoid confusion about what it was I was really asking about.
My question about replacements really asks, in effect, for an "algorithm" for taking a tube and finding a semiconductor circuit with the same characteristics. But the outlines of an algorithm can sometimes emerge from the detailed study of a particular example. So let's make the question of replacements very concrete: the 50LG6GT is a beam power tube. From looking at the tube diagram in the RCA Receiving Tube Manual, p.460, I see that it is a pentode with a 7AC base. Here are the numbers:
Heater Voltage (ac/dc): 50V Heater Current: 0.15A Peak Heater-Cathode Voltage: plus or minus 90 volts, max. Direct interelectrode capacitances (approx): Grid 1 to plate: 0.6 pF Grid 1 to cathode, heater, grid 2 and grid 3: 15 pF Plate to cathode, heater, grid 2 and grid 3: 9.5 pF
Class A Amplifier
----------------- Maximum ratings (Design-Center Values) Plate Voltage: 200 V Grid 2 (screen-grid) voltage: 125 V Plate dissipation: 10 W Grid 2 input: 1.25 W
----------------- Plate supply voltage: 200 V Grid 2 supply voltage: 125 V Grid 1 (control-grid) voltage: --- V Peak AF Grid 1 Voltage: 8.0 V Cathode-Bias resistor: 180 Ohms Zero-Signal Plate Current: 46 mA Maximum-Signal Plate Current: 47 mA Zero-Signal Grid 2 Current: 2.2 mA Maximum-Signal Grid 2 Current: 8.5 mA Plate Resistance (Approx.): 28000 Ohms Transconductance: 8000 micro-mhos Load Resistance: 4000 Ohms Total Harmonic Distortion: 10 percent Maximum-Signal Power Output: 3.8 W
Now, the first step in finding a semiconductor device or circuit to replace this tube is to articulate what characteristics the semiconductor device or circuit has to have. Does anyone know how to do that?
I just did a Google search for semiconductor pentode and found two hits that seem relevant to this thread. One is a US Patent for a MOS semiconductor device having a pentode operation region. The other is a US Patent for semiconductor emulation of vacuum tubes.
The latter sounds like I hit gold. Unfortunately, my ancient browser hangs when I try to look at these sites.
Searching instead for pentode transistor brings up a wikipedia page with a little bit of information.
Other respondents suggested that I look into FETs. I have Paul Richman's book, "MOS Field-Effect Transistors and Integrated Circuits" (1973) and have long wanted an excuse to read it. So, now I have an excuse. But first I want to read the book, "Ions and holes in semiconductors", by William Shockley (1950) because of its historical importance and because it looks like a very good book which will also make it easier to read later books.
I didn't appreciate the remarks of Marra when I first read them:
I just did some more searching for pentode transistor and found the following:
which I just downloaded and looked through. Pritchard seems to be responsible for the hit I mentioned yesterday involving a US Patent for emulating vacuum tubes. The pdf file has comparisons of graphs of responses for tube amplifiers with those for semiconductor amplifiers. It also has an explanation at the end for why it is important to do this: although Eastern European countries still make and sell tubes, that source can be expected to dry up as semiconductors become more common there, so one really does have to try to figure out how to make semiconductor circuits that are as satisfying as the tube circuits. That seems to be what Pritchard is doing.
Pritchard doesn't seem to have any actual tube circuits and seems only to use published data for tubes, whereas he gives his own measurements for his own circuits. Anyway, this article is probably worth studying and I'll try to make sense of it.
Another hit refers to an article by A.G.M.Strollo in the IEEE Transactions on Electron Devices, vol.41 (1994) 616-617, entitled, "Modeling pentode-like characteristics of recessed-gate staticinduction transistor". Here is the abstract: "A new analytical model is described for the pentode-like region of the characteristics of recessed-gate SIT structures. The model allows one to investigate the transition from saturating characteristics of long channel JFET's to nonsaturating behavior of SIT devices, taking into account realistic device geometry."
I don't know what a recessed-gate SIT structure is. However, it sounds as though Strollo is finding it useful to compare at least part of its behavior with the behavior of pentodes. I don't know whether he means pentode tubes but it seems likely.
Here are some updates on my last posting:
Since Pritchard compares graphs, it seems that I should have some way of producing graphs for whatever tube I'm interested in, as well as for whatever semiconductor I might consider replacing it with. That means I need to get interested in modelling the tubes. I have 4 books that might help with this: (a) RCA Receiving Tube Manual (b) Kloeffler's book, Electron Tubes (c) Seely's book, Electron Tube Circuits (d) Radio Amateur's Handbook, 1978
I actually read and worked through (b) a few years ago.
I started reading Shockley and am enjoying it. I also have: (e) The Art of Electronics, 2d edition
It recommends Richman's book, which is reassuring. I think I should try to make whatever use I can of (e) while I do this. I've been aware of the book for a long time, starting with its first edition, and actually built a working 68000 computer system on a breadboard based on its lab book, using borrowed parts and getting lots of help. But I've never had the patience to actually read it. It has the forbidding aspect of 1000 page organic chemistry texts. However, I felt the same way about Frank Gilluwe's book, The Undocumented PC, which is about 1100 pages, until recently when I started trying to learn to write boot sector programs for a PC, and then the book suddenly came to life. So, I'm hoping that the question of this thread, whatever its merits or lack thereof, will catalyze a similar experience with (e). That in itself would make the effort all worthwhile, regardless of the outcome.