that means you measured ~510mVdc across the 68r resistor. Is that correct?
And ~190mVdc across the 27r?
Member
Joined 2009
Paid Member
I have read that it's normal for the VAS and CCS transistors to run hot in this design.
60C is normal for the ZTXs package size in this schematic. If possible mount them closer to the board (shorten their legs as much as possible), so the copper traces can act as heatsink to some degree.
0.5Vdc confirms ~7.4mA through the CCS and that is probably what flows through the VAS.
That gives a dissipation of ~294mW in both devices.
I like to read interesting reports at this forum.
After listening Naim devices at last HI-End Show, I also bought the PCBs NAP140 clone from Ebay (China made from seller xulingmrs ). Little negative that soldering places are very small and there were mistakes in feedback net, R_27k* was disconnected (the feedback resistor was not connected to the base TR2.) Be careful testing PCBs.
I used 3 x BC546, ZTX753, ZTX653, MJE340, MJE350, 2 X 2T819A. 100 pF, 300pF and 470 pF are polystyrene, 200 pF are paper in oil, small caps cyramic, electrolites are tantalum. In PSU blue BC-Vishay 15000uF X 40V.
No coils at output, just 0R22 resistor. Later I'll try coils according to recommendations in forum.
Common ground is on wire between binding posts.
I tryed MJ15024 modern China (fake as I understand later) and RCA 2N3442 from 1980, 2T918A (1981).
As I installed MJ15024 oscillation began and one pair of MJ15024 was burned. I opened the case of burned transistor, crystall size is 2.5mm*2.5mm, it must be 6mm *6mm.
Using 2N3442 oscillation stoped in some positions of wires, I moved wires higher/lover, left/right some santimeters, but the amp was unstable.
The amp is more stable with 2T819A (Soviet TO3 2 mHz transistors 1981 year).
Now the sound of amp is good, very good. Very good is signal/noise value, AMP is very silent. I like it sound, but it's need to heat at least 2-3 minutes. At first seconds the sound is not so good as after heating.
I don't have distortions tester, but good results shows divider (5_OHM/25W and 0.1_OHM in serial). The speaker takes sygnal from 0.1 OHM. Sound is allmost the same clear at full power.
Another very important thing is the schematic of power supply. In report 241 and tnt-audio.com there were recomendations to use torus transformer in power supply with 2 separated secondary windings and 2 bridge rectifier diodes. So am I, my PSU is the same.
In power supply I used 2 bridge rectifier and 2 separated wires at 250W transformer with 500-600W magnetic core.
For the AMPs with single bridge rectifier in power supply unit. If on that AMP's input appears sygnal near 25 Hz or 12.5Hz (for 50 Hz frequency in power supply network) it makes as constant current part in transformer. Next decrease the inductance in magnetic core and increase power consumption of AMP. It's easy to find if You check power consumtion with 25 Hz and 200 Hz test signal. The best transformers are made by 2 wires (secondary windings) simultaniusly, in that case the assymmetry is small and resistance of windings the same. But many times You don't know or not shure about the technology process.
So IMHO, if You have transformer with 2 separated windings make PSU with 2 bridge rectifier, if You use single bridge than carefully test it at 25 Hz. espessially transformers temperature.
The transformers inside NAP200 looks like 500VA-600VA, IMHO 200 VA is quite anough for good bus reproduction for 2*50W AMP with 2 bridge power supply schematic.
The advantages of 2 bridge rectifier schematic was discussed at many forums 10 years ago and I'm surprised why Douglas Self have not mentioned in his bible about this.
The next step to make the AMP better Is to change diodes in PSU, exsept of slow bridges.
I like fast recovery diodes HFA15PB60 (600V/15A) making very good sound. I have not tested CREE's fast Si-carbide diodes that have some advantages. At another forums there are recommendations about C3D20060 600V/20A TO247, C3D10060 600V/10A TO220 and hove to avoid switching noise with RC nets in bridge rectifiers.
Many thanks to NAIM and all enthusiasts, who posted many interestyng ideas hove to improve the schematic and construction.
I plan to order another PCB NAP200 for next experiments and will make small network sound device with Raspberry Pi inside and good USB DAC (with Volumio or OpenELEC sooftware).
After listening Naim devices at last HI-End Show, I also bought the PCBs NAP140 clone from Ebay (China made from seller xulingmrs ). Little negative that soldering places are very small and there were mistakes in feedback net, R_27k* was disconnected (the feedback resistor was not connected to the base TR2.) Be careful testing PCBs.
I used 3 x BC546, ZTX753, ZTX653, MJE340, MJE350, 2 X 2T819A. 100 pF, 300pF and 470 pF are polystyrene, 200 pF are paper in oil, small caps cyramic, electrolites are tantalum. In PSU blue BC-Vishay 15000uF X 40V.
No coils at output, just 0R22 resistor. Later I'll try coils according to recommendations in forum.
Common ground is on wire between binding posts.
I tryed MJ15024 modern China (fake as I understand later) and RCA 2N3442 from 1980, 2T918A (1981).
As I installed MJ15024 oscillation began and one pair of MJ15024 was burned. I opened the case of burned transistor, crystall size is 2.5mm*2.5mm, it must be 6mm *6mm.
Using 2N3442 oscillation stoped in some positions of wires, I moved wires higher/lover, left/right some santimeters, but the amp was unstable.
The amp is more stable with 2T819A (Soviet TO3 2 mHz transistors 1981 year).
Now the sound of amp is good, very good. Very good is signal/noise value, AMP is very silent. I like it sound, but it's need to heat at least 2-3 minutes. At first seconds the sound is not so good as after heating.
I don't have distortions tester, but good results shows divider (5_OHM/25W and 0.1_OHM in serial). The speaker takes sygnal from 0.1 OHM. Sound is allmost the same clear at full power.
Another very important thing is the schematic of power supply. In report 241 and tnt-audio.com there were recomendations to use torus transformer in power supply with 2 separated secondary windings and 2 bridge rectifier diodes. So am I, my PSU is the same.
In power supply I used 2 bridge rectifier and 2 separated wires at 250W transformer with 500-600W magnetic core.
For the AMPs with single bridge rectifier in power supply unit. If on that AMP's input appears sygnal near 25 Hz or 12.5Hz (for 50 Hz frequency in power supply network) it makes as constant current part in transformer. Next decrease the inductance in magnetic core and increase power consumption of AMP. It's easy to find if You check power consumtion with 25 Hz and 200 Hz test signal. The best transformers are made by 2 wires (secondary windings) simultaniusly, in that case the assymmetry is small and resistance of windings the same. But many times You don't know or not shure about the technology process.
So IMHO, if You have transformer with 2 separated windings make PSU with 2 bridge rectifier, if You use single bridge than carefully test it at 25 Hz. espessially transformers temperature.
The transformers inside NAP200 looks like 500VA-600VA, IMHO 200 VA is quite anough for good bus reproduction for 2*50W AMP with 2 bridge power supply schematic.
The advantages of 2 bridge rectifier schematic was discussed at many forums 10 years ago and I'm surprised why Douglas Self have not mentioned in his bible about this.
The next step to make the AMP better Is to change diodes in PSU, exsept of slow bridges.
I like fast recovery diodes HFA15PB60 (600V/15A) making very good sound. I have not tested CREE's fast Si-carbide diodes that have some advantages. At another forums there are recommendations about C3D20060 600V/20A TO247, C3D10060 600V/10A TO220 and hove to avoid switching noise with RC nets in bridge rectifiers.
Many thanks to NAIM and all enthusiasts, who posted many interestyng ideas hove to improve the schematic and construction.
I plan to order another PCB NAP200 for next experiments and will make small network sound device with Raspberry Pi inside and good USB DAC (with Volumio or OpenELEC sooftware).
Hi Dmitry, it seems you have used features of several original models in the one clone - that makes it interesting. It must be pleasing to listen to after all your reading and testing.
You may find that at high power, the low, 0.5A max. current capability of MJE340/350 drivers is not enough for the output transistors. I'm not familiar with 2T819 transistors but most older type power transistors with <4MHz Ft, are also low gain types and that means they need lots of base current when a reactive load like speakers is used. The types fitted to most NAP models are still 4A transistors MJE243/253, though I don't believe the type is critical. I would consider a more suitable driver than MJ340/350 though - perhaps genuine BD139/140 would be better and they're not expensive either.
The design for the NAP200 requires 430VA transformers but 500VA is just a common standard size. Naim amplifiers use larger trafos for sound quality and excess power capability reasons. They are very tough, heavy cranking amps.
Last edited:
Actually BD139/140 are not even close to the TIP41 family or MJE253 type transistors.
If you want sound, use only two types as drivers TIP41 or MJE253... they give best perf., sound differently but way MORE RIGHT then anything else.
For fast KICK and BASS, use FAST transistors on the output. 2SC5200
For some gitarre bass and gigantic room shake... use old: MEV 2N3055 transistors for example.
KEEP ZTX753, ZTX653 in VAS, they are best. And do sound good with lossy tantalums on feedback and bias bypass.
BC546 as input and CCS not so good idea...
Do you have current limit installed, DmitryII ? "VI limit" i mean.
Tell more about naim transformer technology... i am soo interested in this one 😛
If you want sound, use only two types as drivers TIP41 or MJE253... they give best perf., sound differently but way MORE RIGHT then anything else.
For fast KICK and BASS, use FAST transistors on the output. 2SC5200
For some gitarre bass and gigantic room shake... use old: MEV 2N3055 transistors for example.
KEEP ZTX753, ZTX653 in VAS, they are best. And do sound good with lossy tantalums on feedback and bias bypass.
BC546 as input and CCS not so good idea...
Do you have current limit installed, DmitryII ? "VI limit" i mean.
Tell more about naim transformer technology... i am soo interested in this one 😛
Last edited:
Some general advice for experimenters...
Tantalums are very easily destroyed or damaged by reverse polarization. When a circuit is not stable or something blows up on you, you might end up accidentally damaging your tantalum caps. So why not use more robust caps or cap combos as a temporary measure until you are sure the circuit is behaving properly?
The Naim amps are run at the edge of stability. That's why the older NAP250s would start to oscillate unless an appropriate length/type of speak cable was used. To keep things stable, Naim carefully chooses parts and compensation and grounding arrangements. It's like a very fine tuning and it is very very difficult unless you have the equipment and experience. Oscillation will not help your sound quality nor the longevity of your components and speakers.
So why not make things easier when you are getting started and slow the circuit down? Otherwise you are sort of putting rockets on a skateboard. Get it working properly first and later try pushing it to the bleeding edge in a controlled manner.
The easiest place to start is to increase the miller capacitor, say to >100pF. Make sure it is a high frequency capacitor so that it kills the gain of that transistor at very high frequencies. Keep the leads short.
Make sure your grounding arrangements are proper or you can get oscillations this way. Always use separate wires to connect your input ground and Zobel grounds to your star ground.
Make sure your power supply leads are tight to minimize loop inductance - inductance makes the psu caps less effective at high frequencies.
Trust me on this. An amp doesn't have to have silly-high bandwidth to sound good. We can only hear 20kHz! An amp will sound good mostly because its open loop performance is good. Gobs of feedback around a poor circuit will not sound good. A good circuit with low feedback will almost always sound better than a poor circuit with lots of feedback.
So start slow. make the circuit good. Then carefully add more feedback until your circuit sounds worse because of it. 😎
Tantalums are very easily destroyed or damaged by reverse polarization. When a circuit is not stable or something blows up on you, you might end up accidentally damaging your tantalum caps. So why not use more robust caps or cap combos as a temporary measure until you are sure the circuit is behaving properly?
The Naim amps are run at the edge of stability. That's why the older NAP250s would start to oscillate unless an appropriate length/type of speak cable was used. To keep things stable, Naim carefully chooses parts and compensation and grounding arrangements. It's like a very fine tuning and it is very very difficult unless you have the equipment and experience. Oscillation will not help your sound quality nor the longevity of your components and speakers.
So why not make things easier when you are getting started and slow the circuit down? Otherwise you are sort of putting rockets on a skateboard. Get it working properly first and later try pushing it to the bleeding edge in a controlled manner.
The easiest place to start is to increase the miller capacitor, say to >100pF. Make sure it is a high frequency capacitor so that it kills the gain of that transistor at very high frequencies. Keep the leads short.
Make sure your grounding arrangements are proper or you can get oscillations this way. Always use separate wires to connect your input ground and Zobel grounds to your star ground.
Make sure your power supply leads are tight to minimize loop inductance - inductance makes the psu caps less effective at high frequencies.
Trust me on this. An amp doesn't have to have silly-high bandwidth to sound good. We can only hear 20kHz! An amp will sound good mostly because its open loop performance is good. Gobs of feedback around a poor circuit will not sound good. A good circuit with low feedback will almost always sound better than a poor circuit with lots of feedback.
So start slow. make the circuit good. Then carefully add more feedback until your circuit sounds worse because of it. 😎
Sorry for long delay.
Very interesting advices, I'll take in consideration all of them.
Very sensible idea is to test and tune AMP with more robust caps except tantalum, many thanks to traderbam.
To rensil: I didn't install transistors of VI limiter.
About the NAIM transformers, I don't have information about the transformers. Transformer is heavy device, it's not easy to carry it from fare away In Moscow are popular Talema and Torel (from Tula).
Talema are made in Chech Rep. with British steel cores (according to data). Secondary is made by 2 wires simultaneously, so difference in resistance is ~1mOhm and difference in voltages ~0V.
Inductance
500 VA - 1.67 Tl
625 VA - 1.66 Tl
800 VA - 1.65 Tl
1000 VA - 1.63 Tl
Transformers Talema are also good and 1,5~2 times cheaper.
Good advice from Ian Finch, I have BD139, now I'll order BD140 and MJ243/253 for next experiments.
As I find the highest fake statistics deals with transistors in TO-3 case.
In TO-247 case now the most linear and stable (h and capacity with different Ic) from popular transistors are MJ21194, also in comparison to 2SC2922.
If somebody tested the MJ21194 in NAP AMPs, please inform about the results.
Very interesting advices, I'll take in consideration all of them.
Very sensible idea is to test and tune AMP with more robust caps except tantalum, many thanks to traderbam.
To rensil: I didn't install transistors of VI limiter.
About the NAIM transformers, I don't have information about the transformers. Transformer is heavy device, it's not easy to carry it from fare away In Moscow are popular Talema and Torel (from Tula).
Talema are made in Chech Rep. with British steel cores (according to data). Secondary is made by 2 wires simultaneously, so difference in resistance is ~1mOhm and difference in voltages ~0V.
Inductance
500 VA - 1.67 Tl
625 VA - 1.66 Tl
800 VA - 1.65 Tl
1000 VA - 1.63 Tl
Transformers Talema are also good and 1,5~2 times cheaper.
Good advice from Ian Finch, I have BD139, now I'll order BD140 and MJ243/253 for next experiments.
As I find the highest fake statistics deals with transistors in TO-3 case.
In TO-247 case now the most linear and stable (h and capacity with different Ic) from popular transistors are MJ21194, also in comparison to 2SC2922.
If somebody tested the MJ21194 in NAP AMPs, please inform about the results.
MJ21194 is a To3 device.
MJL21194 is a To264 device.
Neither are the same as To247 packaging.
MJL21194 is a To264 device.
Neither are the same as To247 packaging.
Why?
My clone amp Naim is stable.
Why you need a electrical circuit сapacitor 470pF and resistor 680R (сapacitor 470pF and resistor 390R)?
You show the schematic of the original NAP250 there. It has much in common with NAP 200 but none of the components in 250/140 clones will be the same type as specified for either model. Not the semis, caps, or even the larger resistors. This means that even if the amplifier works fine and is stable, it won't sound much like the original product and that is true of many other cloned amplifiers.
What you can't see from looking at the Caowei NAP200 clone shown earlier, is the improved dual slope VI limiter and some minor changes to the IPS current source etc. Including its dual-mono design and layout, the original is improved from a technical perspective but take care that as you make improvements, the sound quality doesn't just get worse, rather than better. Pay attention to what you are playing with here - it is not just like any generic amplifier design. This one exploits distortion rather than eliminating it so making it better with upgraded components and sub-circuits can be and in my experience, usually is counter-productive.
In this thread, many people will be looking to find the original Naim sound and though it is unlikely you will find it with a cheap clone and the random selection of parts types supplied, using the original or equivalent components and circuit can bring us close enough to satisfy me at least. If I wanted to hear the original sound, I would have bought a pre-loved NAP140, as I once did, since the cost was not much more than buying all the parts required for a clone build anyway.
I suggest that throwing in all the best, high quality parts that might be essential to make a modern, state-of-the-art amplifier, is the wrong way to go about building and modifying NAP clones. Their design is basically a good, standard app. note design from 1970 that had then been tweaked for more impressive sound quality, all nearly 50 years ago. That's what you need to replicate, rather than cobbling together a hybrid of old and new design ideas and components with no idea what that does to sound quality.
In case you think that anyone can tweak distortion for improved sound quality, think about the number of people who can do it successfully and I suspect there are very few, globally - like the number of piano tuners, violin makers etc. who begin with good tonal perception, some musical talent and then amass years of experience. Unfortunately, even with the best efforts, I think that the outcome of modifications for most of us, is projects that soon get dropped in favour of something else.
Regarding instability, I have not yet encountered any that wasn't just an assembly, wiring layout error or a silly modification. The only problem area I'm aware of is the dependence on long, inductive speaker cables when other manufacturers had begun to fit output coils, as in the NCC200 clone.
I would be cautious of modifications that were based on experimentation or anecdotes about original Naim products rather than clone builds. The topic here is Ebay clones and modifications really should use them as the reference for whether they may be a good/bad idea. Let's not compare all the apples with oranges.
Came back with better measurements. The other channel working without problem. Where to look for problems ? Thanks!
An externally hosted image should be here but it was not working when we last tested it.
Unfortunaly this is very true, even a VI limiter play's they role in naim sound.
I still don't understand why nap140/250 used dual centertap transformers to make 1 PSU for stereo ? Damping factor improvement ?
To me, the benefits would be the lower impedance on transformer side as windings are in parallel...
Toroidy.pl guys would make it for some bucks, but is it worth it ? That would be the last item to complete the original design😀
How much hum is normal?
If I place my ear close to the speaker drivers, I can hear a hum at roughly the same frequency as my fluorescent lights. It's impossible to hear from anywhere further than 1-2 feet though.
If I place my ear close to the speaker drivers, I can hear a hum at roughly the same frequency as my fluorescent lights. It's impossible to hear from anywhere further than 1-2 feet though.
Actually, a lot of high-end stereo amplifiers are constructed in a so called "dual mono" way. The basic idea is to have 2 completely separate amplifiers in the one case, saving on the expense and clutter of 2 x monoblocs + preamplifier. The NAP140/250 though, just attempt to gain a little extra stereo separation/isolation from using the separate PSUs.
In proper dual mono construction, the only shared circuits between the amplifiers should theoretically be signal ground and protective earth, at the chassis mains earth connection but there are various degrees of "dual mono" design and though Naim's NAP200 power supplies are electrically separate, they share the same transformer primary and iron core which raises the question of just how "mono" the channels are.
In my experience so far with that amp, having powered both channels with a lab. dual supply, also with both on-board supplies and just one, I have to say Naim's approximation does sound noticeably better than in stereo - in fact close to what I'd expect from any dual mono amplifier using separate transformers, and power supplies, which is what we usually find inside the box. I can't say that NAP140 sounds better than 110 for example, but if there is a consistent difference to be heard, then opinions are divided anyway.
By better, I mean the stereo image is more emphatic and also introduces some other surprising stereo effects, that seem to vary between products. This really does make listening more interesting with good speakers but I can live without it, provided the rest of a normal stereo amplifier is up to a good standard of wiring and layout, such that hum and noise is as good or better than typical at < -96db. This is one aspect of dual mono that is not assured by simply dividing a stereo pair into 2. I have had some grounding and layout problems doing this, even by copying a successful commercial build.
Last edited:
There is a little more to using 2 rectifiers too. It was popular in DIY thinking back in the 1970s-80s, where each channel had its own rectifiers and smoothing caps - often to allow the use of cheaper rectifiers without the need for heatsinks. Considering the full 4R output of a NAP 250, that would seem a wise precaution as much as a separation issue.
Normal by what measure? For an audiophile standard amp with nothing connected to the inputs, then none at all. Even with your ear pressing on the woofer...nothing at all. You should expect some feint hiss from the tweeter.