Thanks again Nigel for your prompt and extensive replies.
I have finished soldering in one pcb and I hope to test it when I get the transformer. I will be a 300VA toroid with 2 secondaries x 27-0-27V. Do you think that VA rating is enough?
Another point: I remember reading in this thread to check that the resistance across R17 (marked 1K on my pcb, which I used) should be around 800R for minimum bias. In my case I could only get slightly higher than 600R., no matter how far I turned the trimmer pot. Is this reading correct or am I missing something?
I have finished soldering in one pcb and I hope to test it when I get the transformer. I will be a 300VA toroid with 2 secondaries x 27-0-27V. Do you think that VA rating is enough?
Another point: I remember reading in this thread to check that the resistance across R17 (marked 1K on my pcb, which I used) should be around 800R for minimum bias. In my case I could only get slightly higher than 600R., no matter how far I turned the trimmer pot. Is this reading correct or am I missing something?
Standard SRPP using 6922 tubes.
The gain is a bit on the high side, so I'll probably be changing them a bit. Thinking of a shunt or some sort of resistor based attenuation.
The highest resistance indicates that you have the bias current wound down to min. and that's all you really need to achieve. The resistance may vary from one clone to another so don't be over-concerned that you can't achieve "around" 800R. The purpose is to ensure the pot is set in the minimum bias current position.
The important figure will be the actual bias current you then set when the amplifier is powered.
The rule I use is the VA should be twice the total output power . Be very careful the off load voltage does not exceeded the maximum ratings of the amplifier or capacitors . Allow at least 10% for safety .
In theory a transformer 1/4 this size will just about work . Some low grade designs are built this way . To be clear 2 x 50W requires 2 x 100 VA minimum if running an industrial application ( a motor drive ) in conventional class AB as Naim . Some say the most punishing music is Techno at 1/3 industrial rating . That implies 66 VA would be enough . In practice even the most cost obsessive company would not go that low . I seem to remember the Quad 405 was 300 VA for 2 x 100 W , that was perhaps as far as anyone should go in cost cutting a high quality design . That is 50 % above total output power . Something to remember . Many amps sound OK when giving as much as 10 % distrotion . Sometimes the power output is considerably higher than book specification . That is where the Quad is perhaps a bit small .
Class D for TV's seems to be a different world , very mean . High book power ratings which in practice are never used . If building class D I would still use the minimum recommendations for class AB ( + 50% ) , too much of the spec sheets I suspect is TV based . Transformers are cheap and like having a big engine guarantees results . Too big is perhaps 1000 VA , it brings fuse blowing problems . Even 400VA if the transformer is not the best .
Class A is usually well described in any design . It requires more VA .
Thanks folks for your insightful replies. Regarding the resistance across R17, I reasoned that the 200R difference is due to the difference in the value of R17 - 1K as against 1.2K.
Regarding power transformer VA rating for Class AB, that's a good rule of thumb to keep in mind: 2x the total power output.
Thanks to all who contributed to this site.
Best wishes for this festive period.
Regarding power transformer VA rating for Class AB, that's a good rule of thumb to keep in mind: 2x the total power output.
Thanks to all who contributed to this site.
Best wishes for this festive period.
Hi Sonata . Do you have the circuit with R17 , my circuit only has values .
My power x 2 came from industrial uses in my job . It seems very reliable . Many years ago I bought some ILP modules for driving motors . The unit was 60 W rms 8 ohm with a heat sink I guess of about 1.5 degrees / watt . The load was stated as 16 watts . About 18 months later the modules started to fail . I phoned ILP who said I was being ridiculous to expect a 60 watts amplifier to give continuous 16 watts ( truely ) . Instead of being angry I listened . Ian said ( I of ILP ) 1/6 continuous was an industry standard or sometimes called crest factor . With some class D I note 1/10 ! Techno 1/3 .
I think Ian was being too unkind to his design and he put me on the right track . To survive 18 months proves it to be basically OK . Out of desperation I chopped the modules open to reuse the heat-sinks . I fitted a 150 watt MOSFET unit and never had a problem since .
The 16 watt load turns out to be 28 VA . 28 x 6 does need a 150 Watt amp if using Ian's rule .
My other rule is . If you can not hold a heat sink for 10 seconds minimum it probably is too small . Sometimes when far from home my hands and ears are my only test equipment . Setting bias by hand and ear it perfectly OK . Small driver transistors and regulators can also be hand tested . Voltages must be respected for safety . A small transistor might be at a higher temperature than a heat sink by this method which for complex reasons is OK . The lower thermal conductivity of the case compensates and gives relatively a correct result . Some transistor that still work might reach 100 C ! This is a crude test . It will sometimes spot a problem . A cheap clip on heat sink might save the day . If a bias component it will change results and this must be checked . Audiophiles in the past would power their amplifiers permanently stating better sound . They were right as the bias might take some time to settle due to heat . Keeping the heat constant in the bias circuit helps ( when idling ) .
My power x 2 came from industrial uses in my job . It seems very reliable . Many years ago I bought some ILP modules for driving motors . The unit was 60 W rms 8 ohm with a heat sink I guess of about 1.5 degrees / watt . The load was stated as 16 watts . About 18 months later the modules started to fail . I phoned ILP who said I was being ridiculous to expect a 60 watts amplifier to give continuous 16 watts ( truely ) . Instead of being angry I listened . Ian said ( I of ILP ) 1/6 continuous was an industry standard or sometimes called crest factor . With some class D I note 1/10 ! Techno 1/3 .
I think Ian was being too unkind to his design and he put me on the right track . To survive 18 months proves it to be basically OK . Out of desperation I chopped the modules open to reuse the heat-sinks . I fitted a 150 watt MOSFET unit and never had a problem since .
The 16 watt load turns out to be 28 VA . 28 x 6 does need a 150 Watt amp if using Ian's rule .
My other rule is . If you can not hold a heat sink for 10 seconds minimum it probably is too small . Sometimes when far from home my hands and ears are my only test equipment . Setting bias by hand and ear it perfectly OK . Small driver transistors and regulators can also be hand tested . Voltages must be respected for safety . A small transistor might be at a higher temperature than a heat sink by this method which for complex reasons is OK . The lower thermal conductivity of the case compensates and gives relatively a correct result . Some transistor that still work might reach 100 C ! This is a crude test . It will sometimes spot a problem . A cheap clip on heat sink might save the day . If a bias component it will change results and this must be checked . Audiophiles in the past would power their amplifiers permanently stating better sound . They were right as the bias might take some time to settle due to heat . Keeping the heat constant in the bias circuit helps ( when idling ) .
Last edited:
Hi Nigel,
You are right:- neither the pcb nor the circuit diagram is marked with component identifiers, only values. My pcbs came from hi-fidiy.net around March-April this year. R17 is easily recognisable because it across the bottom pins of the 2K trimmer. On my pcb (and circuit diagram) it is marked 1K whereas on other schematics it is shown as 1k2. My schematic is the same as that in post #19.
Regarding your power supply explanation I don't quite understand. Earlier on you say a VA rating equal to 2x the total power output (for Class AB) is OK and I'm basing my requirements on this rule of thumb (70+70=140) x2 = 280VA (say 300VA). Maybe I'll go for 400.
Also, many thanks for your tips about temperature 'testing'.
Best regards,
Joe A
You are right:- neither the pcb nor the circuit diagram is marked with component identifiers, only values. My pcbs came from hi-fidiy.net around March-April this year. R17 is easily recognisable because it across the bottom pins of the 2K trimmer. On my pcb (and circuit diagram) it is marked 1K whereas on other schematics it is shown as 1k2. My schematic is the same as that in post #19.
Regarding your power supply explanation I don't quite understand. Earlier on you say a VA rating equal to 2x the total power output (for Class AB) is OK and I'm basing my requirements on this rule of thumb (70+70=140) x2 = 280VA (say 300VA). Maybe I'll go for 400.
Also, many thanks for your tips about temperature 'testing'.
Best regards,
Joe A
Hi and season's greetings to all.
70W is only the nominal maximum output power into a resistive 8R load. If you have lower impedance speakers (many speakers labelled 8R in fact now use generic 6R HT speakers) the load impedance can fall as low as 3.5R in the critical mid-bass power range.
The other important factor is the truth about the output transistors supplied. If your kit came with large Sanken MT200 transistors, you can be certain these are fakes. If genuine, they would not be offered at lower prices than those kits offering much cheaper 2SC5200/A1943 transistors. The actual chips inside are smaller, probably substandard chips from other production which will likely work OK until you test them at >20W. I have had dozens of these kits shown to me as faulty and this is mostly the issue. I routinely discard these when I open the packages now. Other components seem true to label in my experience so far.
This, sadly, is the all too common downside of cheap Asian clone kits and semiconductor purchases generally.
In reality, you will never use your amplifier at full power continuously unless you perform sustained maximum power sinewave testing or don't have a volume control.
At home, very few of us use power levels greater than a couple of watts so it does not really matter what size of transformer you fit - It is a matter of matching the amplifier output capacity to meet a manufacturer's continuous maximum power specification that we are talking about. It's unnecessary for enjoyment and expensive but it does make projects look bigger, at least.
400VA for 2x 100W/4R seems excessive to me and you won't notice any loss using a 300VA model. This and a dual mono version using 2 x 160VA transformers are the most popular assemblies. Check out the many pics in older posts here and see what works for other DIYs.
70W is only the nominal maximum output power into a resistive 8R load. If you have lower impedance speakers (many speakers labelled 8R in fact now use generic 6R HT speakers) the load impedance can fall as low as 3.5R in the critical mid-bass power range.
The other important factor is the truth about the output transistors supplied. If your kit came with large Sanken MT200 transistors, you can be certain these are fakes. If genuine, they would not be offered at lower prices than those kits offering much cheaper 2SC5200/A1943 transistors. The actual chips inside are smaller, probably substandard chips from other production which will likely work OK until you test them at >20W. I have had dozens of these kits shown to me as faulty and this is mostly the issue. I routinely discard these when I open the packages now. Other components seem true to label in my experience so far.
This, sadly, is the all too common downside of cheap Asian clone kits and semiconductor purchases generally.
In reality, you will never use your amplifier at full power continuously unless you perform sustained maximum power sinewave testing or don't have a volume control.
At home, very few of us use power levels greater than a couple of watts so it does not really matter what size of transformer you fit - It is a matter of matching the amplifier output capacity to meet a manufacturer's continuous maximum power specification that we are talking about. It's unnecessary for enjoyment and expensive but it does make projects look bigger, at least. 400VA for 2x 100W/4R seems excessive to me and you won't notice any loss using a 300VA model. This and a dual mono version using 2 x 160VA transformers are the most popular assemblies. Check out the many pics in older posts here and see what works for other DIYs.
Thanks Ian for that explanation which leads me to order a 300-350VA toroid. (I don't have much space for large dia trfx.) I also looked at older posts and as you have suggested there this VA rating should be OK for my needs.
I am at the stage where I would like to test my first finished pcb before committing myself to buying the transformer and fitting the whole setup in my enclosure. I still consider myself a newbie wrt solid state projects so I would like to test it with a small (30VA) 15-0-15Vac trafo to avoid major damage/fireworks! I would like to ask:
1. Is this trfx too small and the +/- voltages too low?
2. If the above is OK, what is a reasonable target bias current and DC offset?
3. With such a setup will it be possible to hear some music on cheap efficient speakers just to ensure that it works properly? (I don't have any test equipment)
Many thanks in advance for your comments and advice.
Happy New Year to all DIYA members.
Joe A
I am at the stage where I would like to test my first finished pcb before committing myself to buying the transformer and fitting the whole setup in my enclosure. I still consider myself a newbie wrt solid state projects so I would like to test it with a small (30VA) 15-0-15Vac trafo to avoid major damage/fireworks! I would like to ask:
1. Is this trfx too small and the +/- voltages too low?
2. If the above is OK, what is a reasonable target bias current and DC offset?
3. With such a setup will it be possible to hear some music on cheap efficient speakers just to ensure that it works properly? (I don't have any test equipment)
Many thanks in advance for your comments and advice.
Happy New Year to all DIYA members.
Joe A
Hi,
I saw your post and decided to chime in
The 30VA traf is perfectly fine for testing, as far as you don't run the amp at max power. You can definitely hear music with it. Long time ago I tested mine with +/-20Vdc, 0.5A power supply and I know it's working. With 2x15Vac you'll have about the same voltage - no problem here. The Naim amp is like a big IC and has very wide operational voltages, limited only by transistor power ratings and speaker load.
Target bias current is about, say, 40mA per channel (total for the channel). You can set it by measuring voltage across the emitter resistors at the output - about 4.5-5.0 mV is fine, and the amp does not gain anything with higher bias. If everything is OK, expect DC offset somewhere below 30mV, mostly depending on your input and feedback resistors matching.
Regards and Happy New Year!
Happy new year all!
Hi again
15V transformer windings with a diode bridge will give ~21VDC rail supplies. That will certainly work since the Naim still functions down to ~12V rails, AFAIK. The bias will need to be set to the same specified voltage/current. Note though, that bias control may not be as good as it could be and power needs to be kept low, due mainly to the low current capability, to just a few watts per channel.
Don't allow the transformer to run more than hand warm and use smaller electrolytic capacitors, say 3,300 uF, to prevent damage to the transformer, though I suspect it will not sound quite right in the bass at typical output levels as the supply voltages will be bouncing up and down (sagging) quite severely but without comparison, you obviously can't tell, can you?
I would be concerned about not having even basic test equipment. You absolutely must test the voltages and especially that associated with bias current, in order to set it correctly since it's not an arbitrary setting. It is critical to stability and safety. Should you get it badly wrong, it will be curtains for the amplifier.
Further, you must be able to measure voltages in order to fault-find when components are damaged or you have made an assembly error. This is quite likely, despite best efforts and using recommended/supplied parts and careful wiring. To me, flying blind without even a cheap DMM is just trusting extreme luck. Please, buy or borrow a DMM that can read down to millivolts DC. The measurement required is very small indeed.
Generallly speaking, whenever we decide to go our own way with operating amplifiers, the problems unfold as we proceed and we need some degree of experience and knowledge to succeed. If you are starting out, try to stay within the kit recommendations and standard DIY suggestions here. At least other builders reading will have relevant experience and be able to assist.
Hi again
15V transformer windings with a diode bridge will give ~21VDC rail supplies. That will certainly work since the Naim still functions down to ~12V rails, AFAIK. The bias will need to be set to the same specified voltage/current. Note though, that bias control may not be as good as it could be and power needs to be kept low, due mainly to the low current capability, to just a few watts per channel.
Don't allow the transformer to run more than hand warm and use smaller electrolytic capacitors, say 3,300 uF, to prevent damage to the transformer, though I suspect it will not sound quite right in the bass at typical output levels as the supply voltages will be bouncing up and down (sagging) quite severely but without comparison, you obviously can't tell, can you?
I would be concerned about not having even basic test equipment. You absolutely must test the voltages and especially that associated with bias current, in order to set it correctly since it's not an arbitrary setting. It is critical to stability and safety. Should you get it badly wrong, it will be curtains for the amplifier.
Further, you must be able to measure voltages in order to fault-find when components are damaged or you have made an assembly error. This is quite likely, despite best efforts and using recommended/supplied parts and careful wiring. To me, flying blind without even a cheap DMM is just trusting extreme luck. Please, buy or borrow a DMM that can read down to millivolts DC. The measurement required is very small indeed.
Generallly speaking, whenever we decide to go our own way with operating amplifiers, the problems unfold as we proceed and we need some degree of experience and knowledge to succeed. If you are starting out, try to stay within the kit recommendations and standard DIY suggestions here. At least other builders reading will have relevant experience and be able to assist.
Many, many thanks Ian and Ruwe for your clear and helpful comments/advice.
I've built a few projects in the past, most of them without major problems. My inexperience boils down mainly to lack of technical knowledge in fault-finding (though with valves things are much simpler - I find).
My test equipment is limited to a couple of DMMs, so I can measure basic things. I don't have oscilloscopes, distortion analyzers, freq. generators, etc. The proposed use of the small transformer is basically to test that the module is working properly.
So, besides giving you my best wishes for the New Year, Thanks a lot for your guidance in an otherwise hazy course in getting this project up and running properly.
JA
I've built a few projects in the past, most of them without major problems. My inexperience boils down mainly to lack of technical knowledge in fault-finding (though with valves things are much simpler - I find).
My test equipment is limited to a couple of DMMs, so I can measure basic things. I don't have oscilloscopes, distortion analyzers, freq. generators, etc. The proposed use of the small transformer is basically to test that the module is working properly.
So, besides giving you my best wishes for the New Year, Thanks a lot for your guidance in an otherwise hazy course in getting this project up and running properly.
JA
NAP 140 Clone power supply kit
I am in a process of assembling NAP 140 clone I bought from ebay. I also bought a power supply/speaker protection combo kit (link) and I have a question. I finished assembling power supply kit and it puts out + and - 42VDC but it won't dissipate. It stays charged and I'm wondering how I can bleed it out before touching it for installation. Any suggestions?
I am in a process of assembling NAP 140 clone I bought from ebay. I also bought a power supply/speaker protection combo kit (link) and I have a question. I finished assembling power supply kit and it puts out + and - 42VDC but it won't dissipate. It stays charged and I'm wondering how I can bleed it out before touching it for installation. Any suggestions?
Hi,
Disconnect from the source and just touch across the capacitors with a small value resistor. Anything from 3 to 100 ohms will do the trick. The capacitors will discharge through the resistor, lower value - faster discharge. Higher resistor value would work too, but you'll have to wait more than few seconds... Don't worry about the power rating of the resistor, it won't have the time to overheat anyway. I know some people who like to do it with a screwdriver across the caps, but I find the use of resistor more elegant, unless you're into welding and fireworks
Don't short the charged capacitor. The very high current pulse will damage the capacitor.
Use a sensible value resistor.
Keep a 1k with a pair of insulated croc clips soldered on for this job.
50V across 1k is only 50mA. Max dissipation is only 2.5W but that high value lasts for just a few milliseconds.
10mF & 1k = 10seconds to discharge down to ~37% of start voltage.
50seconds to virtually no charge.
Use a sensible value resistor.
Keep a 1k with a pair of insulated croc clips soldered on for this job.
50V across 1k is only 50mA. Max dissipation is only 2.5W but that high value lasts for just a few milliseconds.
10mF & 1k = 10seconds to discharge down to ~37% of start voltage.
50seconds to virtually no charge.