Replacing rectifiers by other types
By Jac van de Walle
The errors when using tubes, seem always reverse to the complexity of the application. So the more difficult applications have LESS errors than very easy applications. I think this is because with a rectifier circuit it so easy to knock something together which 'works', but with a PP stage, it would go up in smoke when crazy things are done.
The bad news
Mistakenly, the fact that 'it works' is seen by same wizards as proof for good work. Until...an expensive rectifier says good bye after 500 or 1000 hours, and then the trouble begins. Most specially when a Brand-A tube seems to work well, and a Brand-B tube does not. Then, the shouting about Brand-B is very loud. Even in Chorus in the forums when there is a me-too effect.
More bad news
What will people do, when it goes wrong? From my own experience, I can say it is very exceptional, they open their amplifier, do a good measurement, and decide from there, or ask people who know really better.
Most of the time, people are afraid of that, and so they need to take the amplifier to a specialist. Though in fact this is the best thing to do, for many it is the most unpreferred option of all. So the diagnosis quickly gets stuck at this level. Any solution which leaves the amplifier in it's place is considered preferred. Then, if I say, just buy cheap rectifiers, and replace them when they brake down. Most of the time users feel only frustrated. They try to blame the problem to the seller, or the brand, or anywhere else. But they are not stupid, and at some point begin to accept, the problem is inside the amplifier itself. Then the amplifier has lost it's charisma, and it will be sold.
A better way is explained below here.
Things to do
My main advise, when working with an existing amplifier, is to have a good look at the the tube type designation. If the replacement tube has another designation, the probability you can just use it, is low. Not saying it can't be done. Of course you can, but it's exceptional, and there is a long list why it's always a much less good idea as the people in the internet are saying. Specially those writers who hide their identy, they only scratch the surface, and I believe not so quickly what they say.
Here are some things on that list.
HiFi Forums. For some reason, people on a higher level are not welcome there. That is fine if you want to know which mains cable sounds best. The problem there is, there are many around who are not there to help you at all. All they want, is poke into discussions. Really good engineers you find on the older kind of forums, where some are member for 25 years. Not on HiFi babble forums. Search on Google for: 'groups.io electronics design' |
| Heater voltage. If two tubes have the same heater voltage, they MUST have the same heater current too. Reason is, no rectifier heater circuit is stabilized. So the very famous decision to replace 5U4G by 274B, will lead to 5..10% higher heater voltage of the 274B, because it uses only 2A, but the 5U4G transformer winding is made to supply 3A. So at only 2A, heater voltage will rise. This is the typical thing which may work fine initially. Until the 274B heater breaks, or it's emission will go down sooner as expected. Specially when the old 5U4G was working flawlessly for 10 years before, and it is still good, the conclusion the 274B is ''bad quality' is quickly drawn. |
Plate current. vs plate voltage. Some rectifiers are high voltage, low current. Some are low voltage high current. So even when the heater specifications are the same, we have here something else which is not the same. Now, if you have no idea whatsoever what is the plate voltage and plate current of the AMPLIFIER. Without that crucial information no specialist can say something specific about it. Only general, like the following. Tubes are generally speaking high impedance devices. That is certainly so when comparing then with solid state. So one way or another, tubes are always low current high voltage. There is however some trade off in the design phase. If you want to build a tube for higher voltage, this can be done by using the SAME system, and specify it for lower current. Also in reverse, the current may be somewhat higher, if the voltage is lower. We come here in the area of so called derating. It is totally normal.
At the same time, this covers a big error we often see: Maximum voltage and maximum current are never possible simultaneously. So yes, 5U4G can produce for instance 500V DC, but as you can see from the curve, only at 175mA. (This is where the 500V curve ends) . And yes it can do 250mA as RCA specifies, but only at 125 Volts. Please, it is EXTREMELY advised to read the ORIGINAL RCA 5U4G data sheet, all text parts of it. Because in there are many explanations and good reasons why things are the way they are. So the ONE AND ONLY RIGHT answer can only come, when you KNOW the working point in your amplifier. That is the green dot in the graph above, just as an example. So here, at 175mA, 350V. Suppose you want to replace a 5U4G by 274B, this working point must be in the WHITE Area of the 274B tube. Then it will work. Someone who answers that question without this piece of homework, is telling only stupid stuff. Now, with 274B this is not possible at 175mA. It is recommended at 140mA by Western Electric, with capacitor input, 5H choke, and 4uF Cap. So if your 5U4G amplifier runs at or below this, you can do it. And otherwise (that is most of the time) you can not. So now, it becomes clear, there is no work around for having to know the amplifier's working point Then, before replacing there are a few more things to look at, like maximum allowed capacitor load, heater current, peak plate current, peak reverse voltage. So the replacement tube won't spark. In reverse, replacing a 274B by 5U4G does not need any checking of maximum data, because 5U4G exceeds them all. BUT.... this may lead to higher plate voltage of your amplifier, and potential risk on problems resulting from that. There are cases where the higher voltage is really beneficial, and sound improves. So we have another "yes, but..." here. Provided the transformer will supply 5V at 3A, but I do not thinks so, because it supplies 5V at 2A, and at 3A, voltage will be lower. So this has to be checked. (No..., not at the forums, but with a Voltmeter). A separate little transformer of 5V, 3A may solve this. |
Replace an directly heated tube by indirectly heated. Some tube were made like that to serve as replacement for that, so these generally come CLOSE, though not the same, and some kind of surprises may occur for that reason. These tubes have the heater connected to the cathode. In that way, taking the electrical stress away from the gap between the heater and the cathode. (Maximum only 5V across it). So such cathodes can be made with less isolation, and therefore good efficiency. Also, such tubes have no risk of heater breakage, as there is no mechanical stress on the heater. So all in all, not a bad concept. Still a problem is, these tubes have less internal resistance. So output voltage will be some 10...20V higher. If that is what you WANT and what you NEED, go ahead. But if that is not want you want, you can't use them. And if you don't know that, you don't know. So then it may work, or it may reduce lifetime of the other tubes in the amplifier. Such a tube is for instance 5V4GT. It seems to me a replacement for 274B, but it will increase the output voltage 10...20V. Other tubes are 83V, replacing the 83. Or, GZ37 replacing 5U4G. |
