After my success with the 807 screen, i'am going to take a look at the CCS screen. I'am going to drop the voltage by about 40V and see if that helps. I didn't quess that the screen voltage was the most critical of the lot. You live and learn.
Shoog
Shoog
I had a look at the CCS today. I now know exactly why it isn't working as expected. Gary Pimms design requires the g1 to draw down a bit of current(1.5ma min). This current is used to drop the anode voltage to near the required cathode voltage across a resistor. The problem is that the TT21 doesn't draw any g1 current in anything below positive bias. This means that there is no current to drop the voltage, Infact this part of the circuit draws 0.45ma, which with a resistor of 100k generates a voltage drop of 48V, way short of the required 300V. If you try to increase the dropping resistor to use the availble current, say 650K, then the voltage reference becomes starved and were back to the current source switching itself on and off and hence popping and distortion.
Conclusion Gary Pimms basic pentode CCS will not work with a TT21/Kt88 in any useable setup. Ii have toyed with the idea of bleeding off a bit of current with a resistor to earth, but this unfortunately bypasses the effective impedence of the CCS, which sends it from in excess of 3Mohm to a paltry 300K.
Back to the drawing board on this one. Its amazing that the CCS worked at all, and even more amazing that it produced the required 45ma of current.
I will look again at my original design with just two resistors been bootstrapped by the TT21. Wont perform as well, but hopefully it will work.
Shoog
Conclusion Gary Pimms basic pentode CCS will not work with a TT21/Kt88 in any useable setup. Ii have toyed with the idea of bleeding off a bit of current with a resistor to earth, but this unfortunately bypasses the effective impedence of the CCS, which sends it from in excess of 3Mohm to a paltry 300K.
Back to the drawing board on this one. Its amazing that the CCS worked at all, and even more amazing that it produced the required 45ma of current.
I will look again at my original design with just two resistors been bootstrapped by the TT21. Wont perform as well, but hopefully it will work.
Shoog
Hi there,
Good news, things are looking up.
I decided to check the CCS wiring before doing anything else. One of the wire was crossed over between channels. Putting that right helped a lot. If you saw the tight space I have the tag boards are housed in, you would understand how I could make such a simple mistake. I then tried upping the g1 feed resistor value from 200K to 1.2M, this took the voltage drop from 48V to a respectable 217V, which is just about liveable. This leaves the anode of the 807 at 430V (I think) with a current of 42ma for a dissapation of about 20Watts which is within rating.
Seems stable enough and seems to work OK. The voltage drop predicted was 315V, so the voltage reference is running really on the edge.
Now the only major issue I face is ringing in the Output transformers. These are 200VA mains transformers pressed into service as a stop gap until I can get some autotransformers wound. A nice clean signal goes in, and a dirty spikey signal comes out. On quiet acoustic signals its not to bad, but put in a indie band with lots of jangly guitars and its terrible. This could be the biggest challenge to resolve.
Anyone still listening, then please check in or I will stop posting.
Shoog
Good news, things are looking up.
I decided to check the CCS wiring before doing anything else. One of the wire was crossed over between channels. Putting that right helped a lot. If you saw the tight space I have the tag boards are housed in, you would understand how I could make such a simple mistake. I then tried upping the g1 feed resistor value from 200K to 1.2M, this took the voltage drop from 48V to a respectable 217V, which is just about liveable. This leaves the anode of the 807 at 430V (I think) with a current of 42ma for a dissapation of about 20Watts which is within rating.
Seems stable enough and seems to work OK. The voltage drop predicted was 315V, so the voltage reference is running really on the edge.
Now the only major issue I face is ringing in the Output transformers. These are 200VA mains transformers pressed into service as a stop gap until I can get some autotransformers wound. A nice clean signal goes in, and a dirty spikey signal comes out. On quiet acoustic signals its not to bad, but put in a indie band with lots of jangly guitars and its terrible. This could be the biggest challenge to resolve.
Anyone still listening, then please check in or I will stop posting.
Shoog
Hello! Good work on this amp! I think that your mains output transformers are just saturating. Most of them don't like any audio signal at all, and only some of them are good enought to pass just a watt or two of clean audio. Get a real OT and try...
I think you're almost there! How's the sound of this particular configuration?
I think you're almost there! How's the sound of this particular configuration?
Iam not convinced they are saturating, as they have a very high inductance and there is absolutely no DC in them (hence Parafeed).
I think the source of the ringing maybe another winding which is not used and inadequately damped. I will spend some little time on getting this right. I do eventually expect to use some nice toroidal autotransformers for the output, but cost is an issue at the moment.
It is very difficult to say how they sound so far as I have them hooked up to some crappy fullrange test speakers without cabs. I would say though that they do suggest that they will have good bass. They sound very neutral and un-tubey, certainly not the expected SET sound which a lot of people describe (flabby bass and rolled off highs). Even on the test speakers, with far from optimal output transformers, cymbals sound very promising.
Shoog
I think the source of the ringing maybe another winding which is not used and inadequately damped. I will spend some little time on getting this right. I do eventually expect to use some nice toroidal autotransformers for the output, but cost is an issue at the moment.
It is very difficult to say how they sound so far as I have them hooked up to some crappy fullrange test speakers without cabs. I would say though that they do suggest that they will have good bass. They sound very neutral and un-tubey, certainly not the expected SET sound which a lot of people describe (flabby bass and rolled off highs). Even on the test speakers, with far from optimal output transformers, cymbals sound very promising.
Shoog
Good that! Can you explain me the whole thing about using mains transformers for audio output? Which ones to use?
I'm using one of them (220V-9V) for a sigle tube radio project, a SE of a ECL82. But for hi-fi? And why do you want to use autotransformers?
I'm using one of them (220V-9V) for a sigle tube radio project, a SE of a ECL82. But for hi-fi? And why do you want to use autotransformers?
Mains Toroidals have very good bandwidth and so are ideal for the job of output transformers. In order to use any mains transformer in an output situation you have to use parafeed which involves sending the output signal through a capacitor to block DC, in order to make this work the job of DC biasing of the output tube, which is usually through the output transformer, is performed with either a choke or a Constant Current source. Therefore the DC bias and the output signal are dealt with completely seperately. The output transformer is free to work purely on the AC signal.
When you say that a mains transformer wouldn't work in HIFI, I think you are missing one very important issue. SET output transformers are severly compromised by having to deal with DC. This involves creating an air gap. This makes them have to be big. They suffer from poor bass and rolled off highs. A mains transformer (especially toroidal) has good bandwidth because it doesn't have the air gap. I would guess that even a lowly EI mains transformer has the "potential" to sound better than a conventional air gapped SET transformer. Now if money is no object a proper parafeed output transformer should sound even better, but they are frighteningly expensive. I think this is mainly due to there limited marker rather than any additional expense in manufacture. Infact I would guess they are cheaper to make than SET transformers.
An auto transformer will work even better because all it is doing is taking the output voltage of say 300V and converting it down to say 20V. There is no issue of interwinding capacitance so bandwidth should be even better. Again Toroidals work well as output autotransformers. I doubt that I will be able to source any in Europe and so will investigate the option of custom made from India where they still have the skills and the costs are reasonable.
On the subject of my output transformers. I had the signal scoped up and the transformers were generating horrendous harmonics at least as powerful as the fundamentals. I switched the polarity of the secondaries. I placed a 0.47uf across the output of the unused winding. I placed 1nf caps across the two halves of the used windings. Harmonics are largely gone. There is still a bit of an overshoot problem, but no significant ringing.
Sound is now clear and fairly clean. The overshoot is still introducing distortion on the attack of vocals, but its a 100% better than before. Will mess about with zobels now.
Amp sounds really nice. Lovely midrange but without the distorted sound you can get from valves.
Shoog
When you say that a mains transformer wouldn't work in HIFI, I think you are missing one very important issue. SET output transformers are severly compromised by having to deal with DC. This involves creating an air gap. This makes them have to be big. They suffer from poor bass and rolled off highs. A mains transformer (especially toroidal) has good bandwidth because it doesn't have the air gap. I would guess that even a lowly EI mains transformer has the "potential" to sound better than a conventional air gapped SET transformer. Now if money is no object a proper parafeed output transformer should sound even better, but they are frighteningly expensive. I think this is mainly due to there limited marker rather than any additional expense in manufacture. Infact I would guess they are cheaper to make than SET transformers.
An auto transformer will work even better because all it is doing is taking the output voltage of say 300V and converting it down to say 20V. There is no issue of interwinding capacitance so bandwidth should be even better. Again Toroidals work well as output autotransformers. I doubt that I will be able to source any in Europe and so will investigate the option of custom made from India where they still have the skills and the costs are reasonable.
On the subject of my output transformers. I had the signal scoped up and the transformers were generating horrendous harmonics at least as powerful as the fundamentals. I switched the polarity of the secondaries. I placed a 0.47uf across the output of the unused winding. I placed 1nf caps across the two halves of the used windings. Harmonics are largely gone. There is still a bit of an overshoot problem, but no significant ringing.
Sound is now clear and fairly clean. The overshoot is still introducing distortion on the attack of vocals, but its a 100% better than before. Will mess about with zobels now.
Amp sounds really nice. Lovely midrange but without the distorted sound you can get from valves.
Shoog
OOOOHH thank you for your extensive reply! Nice to see that you like what you've built. Keep doing things this way!
One thing: and for PP output? Are the mains transformers right? Should I use toroidals? Is there a method to calculate the proper transformer's power rating when the audio output is known? I mean with tube-oriented trannies, you just take a tranny that's a little bigger of the amp's max output. With mains-oriented trannies? I've heard of multiplying the power by two or three...
I'm interested in PP output... Thank you!
Sizing is guess work on these, but the rule goes something like this. Work out the wattage from the VA rating down to 50hz. To get down to 30hz double that VA rating, and to get down to 15hz double it again. So for a response down to 15hz at 10watts you would need at least 40VA (sounds way to low to me so maybe I got something wrong). My mains transformers are EI and are about 220VA. I heard someone say that they got satisfactory results with 50VA toroidals, but I can't remember how low he went and what wattage.
In theory PP amps should be able to use mains transformers stock as they don't have to cope with DC. However if the output valves aren't perfectly balanced then there will be DC and they will saturate very quickly. EI would cope with this slightly better than toroidals.
I believe toroidals are better for the job because they are more linear and they have wider bandwidth. Hence proper toroidal output transformers are more expensive than EI's.
Good luck and keep us posted.
Shoog
In theory PP amps should be able to use mains transformers stock as they don't have to cope with DC. However if the output valves aren't perfectly balanced then there will be DC and they will saturate very quickly. EI would cope with this slightly better than toroidals.
I believe toroidals are better for the job because they are more linear and they have wider bandwidth. Hence proper toroidal output transformers are more expensive than EI's.
Good luck and keep us posted.
Shoog
Thank you!
To avoid DC offset in the tranny, assuming a fixed bias PP output stage, should I put a bias pot for each tube? And what about PP cathode bias?
If I can get this to work, I have eliminated a major part of the cost of a typical tube amp. I still have to eliminate the power tranny... Someone suggested SMPS supplies, but I'd hate it. Maybe winding your own PT? Or maybe put a 220>6V tranny for the filaments and then 6>220 one to take up the voltage for the B+? Im' just trying to save something about costs, here in Italy it is difficult to find tube-oriented trannies, and they're extremely costly.
Thanks!
To avoid DC offset in the tranny, assuming a fixed bias PP output stage, should I put a bias pot for each tube? And what about PP cathode bias?
If I can get this to work, I have eliminated a major part of the cost of a typical tube amp. I still have to eliminate the power tranny... Someone suggested SMPS supplies, but I'd hate it. Maybe winding your own PT? Or maybe put a 220>6V tranny for the filaments and then 6>220 one to take up the voltage for the B+? Im' just trying to save something about costs, here in Italy it is difficult to find tube-oriented trannies, and they're extremely costly.
Thanks!
I don't have any experience of tube PP amps, so I can't advise much on that. I don't know whether some form of fixed bias would keep DC stable. You'll have to ask elsewhere.
I wouldn't think winding your own power transformers would be that easy. Ebay has some nice power transformers at good prices. Vintage ones will be better built than most of what you could buy for huge amounts of money new. Be careful about what you buy of Ebay as some of those clowns don't know what they have got and are crap at describing things. I ended up with a few single winding transformers when I wanted center taps.
Your idea of using a 6V-250V step up from your raw 6V transformer works. I have used this in my preamplifier. The only thing to consider is that to get a 250-0-250V transformer you will need two transformers after the 250V-6V, as the primary side will only be a single winding.Should work with primary side 0-120V,0-120V power transformer Insertion losses will be higher (multiply the regulation between the first and second transformer - at a guess), and because of this the power supply will be more prone to sag - not an issue in a preamp, but maybe one in a power amp.
Your on the right track. All these things work - but you'll not get good advise as most people haven't tried them. Its a case of experimmenting and seeing what works. Certainly with the output power transformer option it introduces some issues of stability and distortion which have to be addressed, but there not insurmountable. Certainly you are compromising the end result - but if you can't afford the expensive parts, then you weren't going to have any result to compromise anyway.
My feeling is that theres a certain subset of audio freeks who don't think it can sound good if they haven't sunk huge amounts of money into their project. I'am not one of them. My current amp will probably end up costing me about €300-400, which I think is a real bargin.
Shoog
I wouldn't think winding your own power transformers would be that easy. Ebay has some nice power transformers at good prices. Vintage ones will be better built than most of what you could buy for huge amounts of money new. Be careful about what you buy of Ebay as some of those clowns don't know what they have got and are crap at describing things. I ended up with a few single winding transformers when I wanted center taps.
Your idea of using a 6V-250V step up from your raw 6V transformer works. I have used this in my preamplifier. The only thing to consider is that to get a 250-0-250V transformer you will need two transformers after the 250V-6V, as the primary side will only be a single winding.Should work with primary side 0-120V,0-120V power transformer Insertion losses will be higher (multiply the regulation between the first and second transformer - at a guess), and because of this the power supply will be more prone to sag - not an issue in a preamp, but maybe one in a power amp.
Your on the right track. All these things work - but you'll not get good advise as most people haven't tried them. Its a case of experimmenting and seeing what works. Certainly with the output power transformer option it introduces some issues of stability and distortion which have to be addressed, but there not insurmountable. Certainly you are compromising the end result - but if you can't afford the expensive parts, then you weren't going to have any result to compromise anyway.
My feeling is that theres a certain subset of audio freeks who don't think it can sound good if they haven't sunk huge amounts of money into their project. I'am not one of them. My current amp will probably end up costing me about €300-400, which I think is a real bargin.
Shoog
Thanks! About Ebay... those things you mention refer to american objects for sale: their prices are good but shipping to Italy would be expensive. And on italian Ebay you just can't find tube trannies for cheap... the few ones are just new expensive trannies resold.
Well, I guess that using the two trannies option isn't a good idea for power amps... I'm using them on a little one-tube radio, and with HUGE losses, maybe 50%. So I guess I have to buy them... I just find a place in Italy (http://www.elettronicanovarria.it) that seems to have good prices...
I want to use mains outputs trannies for a guitar amp: would those problems with them cause trouble? I mean for a little 20W amp (I'm thinking about EL84 or 6V6 in PP), where I don't need hi-fi sound...
Well, I guess that using the two trannies option isn't a good idea for power amps... I'm using them on a little one-tube radio, and with HUGE losses, maybe 50%. So I guess I have to buy them... I just find a place in Italy (http://www.elettronicanovarria.it) that seems to have good prices...
I want to use mains outputs trannies for a guitar amp: would those problems with them cause trouble? I mean for a little 20W amp (I'm thinking about EL84 or 6V6 in PP), where I don't need hi-fi sound...
Try German Ebay, its really good and German postage is not to expensive. Plenty of old Tube stuff from East Germany.
I was testing the amp tonight to see what problems remain to be solved. The output transformers are saturating with high output, of low bass, which is not what I was expecting. It unusual as it only seems to be saturating on one side of the waveform. I think there must be some other issue which i'am just not getting. I will try to find some references to the other people who have used mains transformers as outputs.
Shoog
I was testing the amp tonight to see what problems remain to be solved. The output transformers are saturating with high output, of low bass, which is not what I was expecting. It unusual as it only seems to be saturating on one side of the waveform. I think there must be some other issue which i'am just not getting. I will try to find some references to the other people who have used mains transformers as outputs.
Shoog
Hi! Thanks for the suggestions... About those trannies, I've read that cheap trannies have some permanent magnetic offset in them that distorts the waveform asimmetrically. And that's a thing that trannies are designed to NOT do... but maybe cheap or defective ones... or just used...
Maybe I'm wrong, I'm starting to think about strange flyback pulses from a trasformer that's not rated for audio output use... you see, inter-winding capacitance, inductance...
Maybe I'm wrong, I'm starting to think about strange flyback pulses from a trasformer that's not rated for audio output use... you see, inter-winding capacitance, inductance...
Tneres overshoot about 5 times the voltage of the signal and this is causing the saturation. I believe the cause is the unused winding. I think it needs a dummy load. Its storing energy and having nowhere to dump it, it collapses and sends a pulse to the secondaries, which is the lowest load path. Unfortunately this will eat up output. Will test tonight and report back.
Shoog
Shoog
Hello! Yes, I think so. It might be the unused secondary. But there are many amps (I'm thinking about PA amps) that have a second secondary (maybe the 100V one, 250ohm) and don't do the things your amp does. Maybe their secondary is just a tap in the only one...
Don't know... I'm not an expert...
Don't know... I'm not an expert...
Placing a 10uf in series with a 47R resistor on the unused secondary has largely eliminated the saturation, but theres still bad distortion which is from ringing. Its gona be tough, and it just might not happen.
Will keep you posted.
Shoog
Will keep you posted.
Shoog
Did some messing tonight. Got the overshoot down to a relatively small spike. Signal is still really distorted. I'am thinking i'am missing something here. The bass sounds OK, and the treble sounds OK, it the midrange which has the issues. I'am wondering if the Parafeed cap is just to small and the cap is saturating and washing the detail away.
When I scope up the input signal, it looks very similar to the output signal, but when I scope up the parafeed cap its lacking in most of the detail I see on the other two readings.
What is the sound of an undersized parafeed cap ?
Shoog
When I scope up the input signal, it looks very similar to the output signal, but when I scope up the parafeed cap its lacking in most of the detail I see on the other two readings.
What is the sound of an undersized parafeed cap ?
Shoog
Made some real progress today. I was doing some measurements with my scope and discovered that my amplifier was infact putting out half the voltage that was going in.
This can't be the way the RH807 was designed, its probably got something to do with the fact that a CCS loaded parafeed amp can't output voltages greater than the supply rail
were as a normal SE can, hence double output. I therefore concluded that all the gain was all been gobbled up in feedback. The 807 would also be achieving greater gain because of the CCS, so this would be making matters worse.
I therefore decided to try reducing the feedback by changing the frb resistor from 100K to 470K. This will mean that it will work more like a pentode, but it should increase the output impedence.
In this case that is not such a bad thing, as I believe the interaction between the low output impedence, the parafeed cap and the primary inductance was causing a resonant peak, hence the distortion in the midrange.
So having done that, (and altered the CCS again because the plate of the 807 had risen to 550V, I brought it back down to 450V). The theory seems to have worked. More useful output with no significant distortion.
The damping of the used secondary has eliminated the overshoot, but since it is in parallel with the output secondary it is loading down the output tube more than I would like. I will look at altering it again.
It came as quite a shock when I finally got clean music out. My conclusion is that the output impedence was to low for parafeed, and the transformer load was wrong. Also with the amount of partical feedback been used the output tube was probably overloading. Sounds quite a bit brighter than before, but if there was a resonance issue then that would be expected.
This can't be the way the RH807 was designed, its probably got something to do with the fact that a CCS loaded parafeed amp can't output voltages greater than the supply rail
were as a normal SE can, hence double output. I therefore concluded that all the gain was all been gobbled up in feedback. The 807 would also be achieving greater gain because of the CCS, so this would be making matters worse.
I therefore decided to try reducing the feedback by changing the frb resistor from 100K to 470K. This will mean that it will work more like a pentode, but it should increase the output impedence.
In this case that is not such a bad thing, as I believe the interaction between the low output impedence, the parafeed cap and the primary inductance was causing a resonant peak, hence the distortion in the midrange.
So having done that, (and altered the CCS again because the plate of the 807 had risen to 550V, I brought it back down to 450V). The theory seems to have worked. More useful output with no significant distortion.
The damping of the used secondary has eliminated the overshoot, but since it is in parallel with the output secondary it is loading down the output tube more than I would like. I will look at altering it again.
It came as quite a shock when I finally got clean music out. My conclusion is that the output impedence was to low for parafeed, and the transformer load was wrong. Also with the amount of partical feedback been used the output tube was probably overloading. Sounds quite a bit brighter than before, but if there was a resonance issue then that would be expected.
- Status
- Not open for further replies.