The rectifier circuit I indicated that may be reasonable is the bridging option on page 45 of linked doc:
https://dalmura.com.au/static/Power supply issues for tube amps.pdf
There is doubt as to the actual datasheet ratings for your 5U4C tube, so you are in a twilight zone by going down the cheap kit amp path. As I read the specs, that tube seemed to have quite low current ratings, which suggests that paralleling the anodes may be the only way to achieve a practical design given your likely B+ loading.
By the way, page 26 of that link gives some background on inserting a valid power transformer winding resistance in PSUD2. I'd be thinking your modelled value of 37 ohm in post 39 is still too low. Also a 150uF ecap is going to have way lower resistance than 5 ohm. This is all quite a detailed topic, with lots of learning curve, so imho keep an open mind and perhaps don't set any circuitry or parts in concrete until you have the actual kit and have made some initial measurements.
https://dalmura.com.au/static/Power supply issues for tube amps.pdf
There is doubt as to the actual datasheet ratings for your 5U4C tube, so you are in a twilight zone by going down the cheap kit amp path. As I read the specs, that tube seemed to have quite low current ratings, which suggests that paralleling the anodes may be the only way to achieve a practical design given your likely B+ loading.
By the way, page 26 of that link gives some background on inserting a valid power transformer winding resistance in PSUD2. I'd be thinking your modelled value of 37 ohm in post 39 is still too low. Also a 150uF ecap is going to have way lower resistance than 5 ohm. This is all quite a detailed topic, with lots of learning curve, so imho keep an open mind and perhaps don't set any circuitry or parts in concrete until you have the actual kit and have made some initial measurements.
Apparently someone with no knowledge has changed the schematic a few times over the years. The original design I found called for a 5Z3P tube not 5U4C, a 47uF cap not 150uF, and finally a 150uF (or 330uF) cap. That produces 298V in PSUdesigner, if I have done it correctly, close to the 310V specified. However, 47uF is still higher than the 32uF max recommended.
A later version shows a more complicated design like this. It appears to show that they acknowledged the output of the transformer would be about 320V instead of 300V when running on 120VAC instead of 110VAC, inserting 10R to get rid of some voltage? The other values shown don't work though, so something went wrong.
If I can figure out what the tube rectifier circuit is supposed to be, then I can move on to imitating it with diodes. I think maybe the bottom one of the two designs above is the intended physical layout to run on 120VAC but with the wrong values.
That was a default value that loaded somehow. Suggested test value?
I won't. I'm just doing my homework and learning. This is my first time looking at tube rectification and also SE. I have time now because it's a holiday week and I may not have so much time later. A lot of people probably don't do their homework and end up with a bad result. I may anyway. 😉 If/when I build it, I am guessing it will be after the first of the year.
A later version shows a more complicated design like this. It appears to show that they acknowledged the output of the transformer would be about 320V instead of 300V when running on 120VAC instead of 110VAC, inserting 10R to get rid of some voltage? The other values shown don't work though, so something went wrong.
If I can figure out what the tube rectifier circuit is supposed to be, then I can move on to imitating it with diodes. I think maybe the bottom one of the two designs above is the intended physical layout to run on 120VAC but with the wrong values.
That was a default value that loaded somehow. Suggested test value?
I won't. I'm just doing my homework and learning. This is my first time looking at tube rectification and also SE. I have time now because it's a holiday week and I may not have so much time later. A lot of people probably don't do their homework and end up with a bad result. I may anyway. 😉 If/when I build it, I am guessing it will be after the first of the year.
Last edited:
It seems that most people build first and ask questions later. I ask questions first and build later. Given the number of these Chinese EL34 single-ended amplifiers floating around, we will be doing a public service if we can come up with a current version of the Chinese SE EL35 amplifier that is proven to work well and reliably with minimal modifications. It's cheap and it's fun, so if it works too, all the better. It doesn't have to be perfect.
Stephe said in another thread to use 5Ц4С with 4.7uF and 150uF, and a 12V bucking transformer, plus add a bleeder resistor for saftey. Great, BUT the bucking transformer is a real PITA if it can't be located inside the chassis, and I don't know the dimensions yet.
Instead of 5Ц4С, the 5Z3 is the original design on the early Chinese schematics, before the design got corrupted, and 10uF is specified on the 5Z3 spec sheets (I think).
So why not just do one of these two options? Wouldn't either one work? Problems with each one? I stuck a 25 Ohm in with the 5Z3 to get rid of a few extra volts. Not sure how many watts would be dissipated by the 200 Ohms in the solid state option. Am I in the ballpark with either option?
Stephe said in another thread to use 5Ц4С with 4.7uF and 150uF, and a 12V bucking transformer, plus add a bleeder resistor for saftey. Great, BUT the bucking transformer is a real PITA if it can't be located inside the chassis, and I don't know the dimensions yet.
Instead of 5Ц4С, the 5Z3 is the original design on the early Chinese schematics, before the design got corrupted, and 10uF is specified on the 5Z3 spec sheets (I think).
So why not just do one of these two options? Wouldn't either one work? Problems with each one? I stuck a 25 Ohm in with the 5Z3 to get rid of a few extra volts. Not sure how many watts would be dissipated by the 200 Ohms in the solid state option. Am I in the ballpark with either option?
Last edited:
A bit of OT:
Many years ago there was a music band in Cuba named "The 5U4", and when asked why the name the founder answered "Because those tubes were always failing and we were always struggling to get new ones".
Many years ago there was a music band in Cuba named "The 5U4", and when asked why the name the founder answered "Because those tubes were always failing and we were always struggling to get new ones".
I was going to say that "The Rectifiers" would be a good band name, but there is a band with that name already!
So far no answers on the possible solutions in post #43. Hmmmmm ... slow day.
Solid state rectifiers are a way better choice, and there is another way to fix the over-voltage situation: place one or more zener diodes where the 200 Ohm resistor (R1) is. You can drop about any voltage to like as long as you get the power dissipation right.
A valve rectifier has datasheet limits on the peak current during operation. The datasheet in post #1 indicates that is 350mA for the 5U4C. In your 5Z3 PSUD2 sim, the continuous peak is circa 550mApk, with a transient hot-start peak of circa 1.5Apk. 5Z3 datasheet indicates a 675mApk rating, but no rating for transient level (although that can sort of be deduced but is arduous).
Note that the 5U4C is likely operating within spec when the input cap value is 2uF and supply resistance is only the 37 ohm (ie. no added 25 ohm). It will depend a lot on the power transformer effective resistance, which you don't know yet, so it isn't appropriate imho to malign the use of the 5U4C if the total B+ current is not above 125mA, which may be reasonable even for a stereo configuration as the schematic shows 2.5mA per channel preamp, so with 60mA per channel output stage then makes 125mA total nominal.
Note that the 5U4C is likely operating within spec when the input cap value is 2uF and supply resistance is only the 37 ohm (ie. no added 25 ohm). It will depend a lot on the power transformer effective resistance, which you don't know yet, so it isn't appropriate imho to malign the use of the 5U4C if the total B+ current is not above 125mA, which may be reasonable even for a stereo configuration as the schematic shows 2.5mA per channel preamp, so with 60mA per channel output stage then makes 125mA total nominal.
Interesting. How do I calculate what Zener diodes to use and how many? Do they dissipate heat? That power has to go somewhere, right? Pardon my newness if I ask silly questions. I had no idea this might work. There must be some downside as nothing is free and everything is a tradeoff somehow.
@trobbins, thanks for that excellent information. Until two days ago, I had no idea how a tube rectifier works. I have learned so much in the past few days that it is incredible. That's why I jumped into this. Today I researched rectifier tubes. I found quotes along these lines:
Chinese 5Z3P tube is the 5U4 tube type (5U4, 5U4G, 5U4GB).
The 5U4 is directly heated. Be aware of the large voltage surge that will occur when the 5U4 is not conducting. This voltage surge may be 50% more than the steady state voltage. So instead of your power supply caps having 425VDC they will be hit with a surge of 600-700 volts for 20-30 seconds before dropping down to the operating voltage when the other tubes warm up and pull down the supply. If the amp was ever turned on with rectifier but without power tubes, that same unloaded high voltage will be present so the caps should be set up to take that.
So why use these tubes (as in the original design before it was corrupted) if we get the instant-on of diodes but all of the disadvantages of having another tube to first wear down in performance and then be replaced? I thought one of the primary benefits of tube rectification was the "slow start" that indirectly heated rectifier tubes provide.
Now they are substituting 5Ц4С, which is indirectly heated, instead of the directly heated Chinese 5Z3P that was in the original design.
My objection to the 5Ц4С tube was based on what appears to be a short rated life on the spec sheet, and there are complaints online about a short lifespan of half a year. Then again they are using two 150uF caps which I now have learned is totally inappropriate for the tube.
I also have read that in a properly designed power supply, a decent 5U4GB, which is equivalent to the Chinese 5Z3P tube from the original design before it was corrupted, can last many years, even in a TV that is on many hours a day. It makes sense because we had tube TVs growing up and they rarely blew a rectifier tube.
I have not ruled out the 5Ц4С now that I know a lot more, but I don't see why using it would be an advantage over solid state unless solid state just can't be done without a different power transformer ($120 USD) or there is some sonic improvement from tube rectification. Some people swear by tube rectification for the "best" sound quality. It just seems inefficient on paper to me. Then there is the "slow turn on" of indirectly heated rectifier tubes.
I'm learning. The key to a successful outcome is to learn before building, not after, like we see so often. 🙂 Also to expect the very real possibility that the project ultimately will go up in smoke as you learn. 😱
Last edited:
As you identify, the initial B+ voltage from using ss diodes or directly heated tube diodes is a concern, but that can be managed - the devil is in the detail, and so until the design is settled it may or may not be a concern.
Imho, objecting to technical and life details without due regard for how and why just continues to promote mis-information. There are many facets to why a valve may fail, but sadly some then pin their feelings to a post comment, that then gets reposted .... One example is rectifier tube arcing, and that rectifier make-model then being lambasted as inferior and should never be used by anyone. Rectifier valves do often arc at the end of their life, and some tend to arc sooner, but for more than a decade it has been known, and more so in recent years due to forums, that adding an ss diode in series with each anode of a rectifier diode avoids arcing. So those who consider certain rectifiers to be poor quality, or complain about short life, really need to spend more time mitigating the issue for starters and moving on.
Imho, objecting to technical and life details without due regard for how and why just continues to promote mis-information. There are many facets to why a valve may fail, but sadly some then pin their feelings to a post comment, that then gets reposted .... One example is rectifier tube arcing, and that rectifier make-model then being lambasted as inferior and should never be used by anyone. Rectifier valves do often arc at the end of their life, and some tend to arc sooner, but for more than a decade it has been known, and more so in recent years due to forums, that adding an ss diode in series with each anode of a rectifier diode avoids arcing. So those who consider certain rectifiers to be poor quality, or complain about short life, really need to spend more time mitigating the issue for starters and moving on.
@trobbins Yes, there is much opinion and bias about most hi-fi topics out there online not backed up by facts, and before it was online it filled printed stereo magazines. That's the nature of free speech and free press and we hope that people will sort out what is accurate and what is not accurate. Fortunately, I am not one of those people who makes decisions without any facts, as evidenced by the fact that I am sitting here trying to learn both SS and tube rectification so I can make an informed decision. My initial impression was that the 5Ц4С was a cheapo Chinese solution that was not ideal. Now I am not so sure after only a few days of looking into the matter. Since I have not yet built yet, I can move towards whatever may be ideal in reality.
So, what are the benefits of tube rectification, both in the case of direct-heated and indirect-heated? Some people swear by indirectly-heated for tube life of the other tubes, and some say tube rectification provides superior sound quality. Are there any facts supporting tube rectification over solid state, especially with "instant-on" directed-heated rectifiers? I have no opinion yet as I have no facts yet, just a bunch of opinions that I have read over the years. My other three tube amps are solid state. That's just the way they were designed.
Last edited:
The first rule of superior sound quality is that there are no rules - it is a subjective topic - people make there own 'rules' up.
Each technology, whether it be valve or ss rectifiers, has its own set of engineering and commercial and ... considerations. If someone chooses lowest purchase cost as their 'rule' for choosing, or key benefit for them, then that's fine, but be mindful that it is their rule or key benefit, and others may view it or choose differently. So I don't think you will unearth rules about what is better or not, but rather more insight as to how to implement a particular technology (ie. the engineering/design aspect).
Each technology, whether it be valve or ss rectifiers, has its own set of engineering and commercial and ... considerations. If someone chooses lowest purchase cost as their 'rule' for choosing, or key benefit for them, then that's fine, but be mindful that it is their rule or key benefit, and others may view it or choose differently. So I don't think you will unearth rules about what is better or not, but rather more insight as to how to implement a particular technology (ie. the engineering/design aspect).
Let me try to put together some facts, without arguing which method is better, at least those that make sense to me and can be measured:
- Tube rectifiers are not efficient. A lot of power is lost in the heater, plate dissipation and the extra high voltage the power transformer needs to supply.
- Tube rectifiers give a soft start to the B+ voltage, which is sometimes desirable.
- Tube rectifiers generate less noise than typical SS diodes, because they have a higher equivalent resistance. SS diodes on/off very quickly, resulting in big current swings, generating ringing noise. There are newer SS diodes that minimise this, also "helpers" like a capacitor and resistor in parallel with diodes, acting as damping circuits.
- Tube rectifiers have a higher output impedance, therefore B+ sags in push-pull amplifiers when you increase the output power. In some applications, like guitar amplifiers, this is often considered an advantage.
- SS rectifiers will last virtually forever if properly selected and operated in the appropriated range. Tubes, by design, do not last that long, especially modern rectifier ones, there seems to be quality issues with current production, at least based on what I read on forums.
- Good tube rectifiers are quite costly. Decent SS diodes like UF4007 cost next to nothing.
Personally, I only use SS rectifiers.
The Zener approach improves the regulation a bit - if that matters. Below is the ckt I came up with to allow comparison
I totally forgot that I have a circuit simulator. I don't remember how to use it though. In your example, what is the output voltage, which appears to be BUS2, when input is 320-0-320? How much power dissipation and heat is generated by the Zener diodes?
Last edited:
The output voltage shown on the graph is based upon your shared simulations. The diode losses total ~9W (all three), whereas the 200 Ohm resistor dissipates ~11W.
I see. Thanks. It's always a tradeoff. Save 2 watts but use three Zener diodes instead of one resistor.
Next question for everyone before I proceed with the power supply decisions, is 310V the ideal voltage for this amplifier? It seems like that's a little low compared to some others. Although tube life always is a consideration, I do want to get as much power from each EL34 as is reasonably possible. The fact that it only uses one output tube per channel makes it a lot more palatable for me to replace tubes compared to having to buy 4 tubes for a push-pull amp.
Next question for everyone before I proceed with the power supply decisions, is 310V the ideal voltage for this amplifier? It seems like that's a little low compared to some others. Although tube life always is a consideration, I do want to get as much power from each EL34 as is reasonably possible. The fact that it only uses one output tube per channel makes it a lot more palatable for me to replace tubes compared to having to buy 4 tubes for a push-pull amp.
Last edited:
The EL34 is biased quite cold at 16.6W. The circuit could handle higher B+, but you may need to increase the 300 ohm cathode resistor to keep EL34 dissipation in check, up to ~22W. It will also depend on the maximum DC current allowed by the output transformer.
Last edited:
That is the primary concern (I made a pun!). 🙂
We have to assume that the output transformers in these extremely inexpensive kits are not very robust. Perhaps my expectations will be exceeded. The PP amp kit from "Huaji Audio" turned out great.
Last edited:
You can keep 60mA at higher B+ by increasing the cathode resistor. It is often quite difficult to obtains detailed OPT parameters that come in cheap Aliexpress kits, unfortunately.