If one is building a Naim clone, then one cannot improve on it. One strives to match it.
If one chooses to use a different transistor then the matching is thrown out of the pram. It will no longer be a clone.
Similarly if one chooses a different current through a stage, then one ends up with something that is not a clone.
If one adds components, for example emitter degeneration resistor/s, then again it makes it not a Naim clone.
A good clone is one that matches the performance of the original.
If one chooses to use a different transistor then the matching is thrown out of the pram. It will no longer be a clone.
Similarly if one chooses a different current through a stage, then one ends up with something that is not a clone.
If one adds components, for example emitter degeneration resistor/s, then again it makes it not a Naim clone.
A good clone is one that matches the performance of the original.
In the ideal world you might use this transistor ( most I bought are fakes ) . It might still need the Vbe shifted up to get the ideal compromise. This device is centre colector so needs the pins sleeving. Alas the world no longer has these ideal devices. The Vbe is still the bigger deal. It can be set with an emitter resistor of typically 10R. This should raise the second harmonic in Naim fashion.
KSA916YTA on paper is very similar to the ones supplied with the kit. Both have very high capacitance. The Y selection on paper might offer higher gain. It also is centre collector. The 40 pF is worst case.
The fact the past had better devices might surprise people. It did.
http://datasheet.octopart.com/2SB716-Hitachi-datasheet-105958.pdf
KSA916YTA on paper is very similar to the ones supplied with the kit. Both have very high capacitance. The Y selection on paper might offer higher gain. It also is centre collector. The 40 pF is worst case.
The fact the past had better devices might surprise people. It did.
http://datasheet.octopart.com/2SB716-Hitachi-datasheet-105958.pdf
Andrew. I have very great doubts that anything about these kits is much like the original. I had some test data alas lost which says the way this kit behaves is nothing like a Naim amp. It makes the whole imbalanced long tail pair idea a feeble joke as the Vbe when at 0.55V stops the trick from working. To be honest if that's the best we can do it needs redesigning to resemble a Denon PMA 250 or whatever.
One thought that comes to mind. After the real designer left Naim it might have been that the " wrong " VAS Vbe became part of a genuine Naim . I have no idea. I do remember a test of NAP 250 that suggested the idea was in use at that point in history. The spectrum was 2nd and 3rd near identical and 4th and 5th lower and equal ( like two staires ). Not ideal. Better than most.
I got very upset with a certain Mr G who makes turntables when he said Mr V wouldn't know what people assume he knew. I now wonder if Mr G was right.
One thought that comes to mind. After the real designer left Naim it might have been that the " wrong " VAS Vbe became part of a genuine Naim . I have no idea. I do remember a test of NAP 250 that suggested the idea was in use at that point in history. The spectrum was 2nd and 3rd near identical and 4th and 5th lower and equal ( like two staires ). Not ideal. Better than most.
I got very upset with a certain Mr G who makes turntables when he said Mr V wouldn't know what people assume he knew. I now wonder if Mr G was right.
As I said my setup is not mounted into a chassis, and there is a lot of jumpers and wires around, it is far from optimal. Just by grounding the metal plate where the setup is mounted, and by changing the jumpers, I got 1.4mvac of noise.
I expect much better figures once I'm into a chassis and with tidier wiring.
I expect much better figures once I'm into a chassis and with tidier wiring.
I'm away from home at the moment, but when I'm back at the weekend I'll test the noise level of my unit that has been in use now for about a month and is in a suitable enclosure etc...
Chris
Nigel, are you referring specifically to the Caowei NAP200 clone here?
Thanks
Chris
Well, let me add to the query. TR4 is the VAS transistor and also a PNP. Still, I can't see the point of messing about with it when ZTX753 is still in production, widely available and has some unique qualities essential to Naim sound. I have to assume Nigel is more about experimenting along the lines of his own pipe dreams and creating something different, as AndrewT reminds us.
Not really and not wanting to say anything negative. Mostly what I am saying is try if you can to have some basic spectrum analyser and by either transistor type or TR4 emitter resistor ( 5R6 or 10 R at a guess ) force the 12 K TR2 resistor into action. This seems to be a common error which time has created. Simply put, there is no reason for that resistor to exist if the balance is withing 15%. Mostly what I am saying is if really wanting what people say they want, someone somewhere must adjust the TR4 to work as designed. I suspect even at Naim it has been a long time since it did. The early Naims were very sweet sounding. I put this down to BDY 56. On reflection a 10 MHz device shouldn't do that ( 1MHz ones perhaps ). Maybe the first ones were better do the LTP balance???? NC200 has advantages if seeing where this goes. I do agree with Andrew we too easilly can loose all contact with the Naim design as was. This one thing might be the excetion. NC200 has more loop gain and we can loose a little to good advantage and not be worse that the original. I suspect the TR4 base needs lifting 0.1 to 0.2 V to circa 0.7V.
I have had a very full on project for work. Good outcomes ( very ). Sorry if my posts are even worse then usual.
Why?
The currents in the LTP balance well with the ~0.5V BE voltage of a ZTX run hot.
Naim (and Avondale) run with balanced currents in each half of the LTP, but choose to generate more 2HD by adding the TR2 collector resistor.
I understand what you mean. It's close enough to think it balanced. That's the point. Somehow the point of the 12 K ( TR2 collector ) has been lost over time ( or the NC200 value ). In a balanced version it should be 0R and then the other resistor set to suit. I think someone said 1.85/1.7 mA NC200. That's very near to balance. I suspect a real 1978 Naim would be at least 65% 35 %. That needs the Vbe of TR4 to be circa 0.65V.
The NCC tail current is ~1mA and it is shared pretty close to 50/50 by each side of the LTP. The Naim values achieve the same current balance albeit in the NAP200 they have increased the tail current quite a bit.
It seems to me that there are two ways to increase the 2HD:
1) Dissimilar currents in each side of the LTP
2) Dissimilar gain in each side of the LTP (effected by the TR2 collector resistor)
I expect you could do both (or they might (partially) cancel in some circumstances) but I think Naim and Avondale choose to do only #2.
It seems to me that there are two ways to increase the 2HD:
1) Dissimilar currents in each side of the LTP
2) Dissimilar gain in each side of the LTP (effected by the TR2 collector resistor)
I expect you could do both (or they might (partially) cancel in some circumstances) but I think Naim and Avondale choose to do only #2.
It's worth a try and should only bring positive things.
One could say the same about the Cdom ( TR4 collector to base ). This is 100 % true and 100 % risky. In theory if the 10R added the capacitor can be reduced to regain HF loop gain. It needs an analyser to be sure ( I would use a cheap speaker as load and possibly kill it to test at 5 watt 10 kHz, 6 inch full range for cars @ 4R ). This transistor claims high capacitiance. If the PCB is a good layout it might not need one at all! Rather you than me. In my work I have to do this. At home when my money I prefer not to.
If there is a gain mismatch in the transistors I doubt is matters much or helps. If it was 360 and 330 for example.
One could say the same about the Cdom ( TR4 collector to base ). This is 100 % true and 100 % risky. In theory if the 10R added the capacitor can be reduced to regain HF loop gain. It needs an analyser to be sure ( I would use a cheap speaker as load and possibly kill it to test at 5 watt 10 kHz, 6 inch full range for cars @ 4R ). This transistor claims high capacitiance. If the PCB is a good layout it might not need one at all! Rather you than me. In my work I have to do this. At home when my money I prefer not to.
If there is a gain mismatch in the transistors I doubt is matters much or helps. If it was 360 and 330 for example.
Do i have a right to post PCB design, copied from NAP250/NAP40 schematic? Its small, i have just verified it is stable +/- 30V laboratory supply, no noise and very DIY friendly... Both, SMT and THT tantalum caps are possible. Lots of groundplanes too. Made in Eagle CAD.
Imo its a such a waste of money and time to buy CHINESE clones and then modify them forever.
Imo its a such a waste of money and time to buy CHINESE clones and then modify them forever.
There's no rule saying you can't apart from the topic actually being about Ebay clones. There are no other recent threads on DIY SMT clones of old designs like the NAP series so why not open a new thread?
It should be an eye-opener for those who are confused by untested opinions of what might work and and what actually does work as a good clone, like replacements for obsolete through-hole components. Whether it is similar in performance and sound quality is something else. It should also be interesting if you source some appropriate equivalents like FZT653/753 for the voltage amplifier stage so that any comparisons are sensible.
Anyway, congratulations on finishing your SMT board(s) successfully
. You obviously have the equipment, talent and time to do this and it would be great to see (or rather hear) the results. However, it's a matter of new skills, equipment and parts stocks if you think SMT will be economical for occasional DIYs. I think that for beginners or occasional DIYs, the basic Chinese or Ebay kits really are much cheaper and more certain than full DIY in SMT or THT format. Few will be prepared to go through a long CAD learning process only for building a few clones which many people can do successfully without much tuition.
BTW folks, some more eye candy: Here is another interesting Caowei kit that hasn't been posted here before. Again, this the deluxe, real deal NAP140 and apparently with correct parts list and at least with hints of using the correct 24-0-24VAC transformer for +/- 34V rails. Finally, an Ebay clone that doesn't just follow the generic NAP 250 design, which was actually designed for regulated 40V rails and bridged mode operation. The first guy to buy is honour bound to post the LTP resistor values and put an end to the guesswork and prognostications here - OK?
DIY KIT NEW 140 Amplifier KIT Power Supply KIT Base ON Naim NAP140 | eBay
The first guy to buy is honour bound to post the LTP resistor values and put an end to the guesswork and prognostications here - OK?
DIY KIT NEW 140 Amplifier KIT Power Supply KIT Base ON Naim NAP140 | eBay
Looks very much like 1K and 21K on the images of the populated PCB. Same as on many images out there of genuine NAP140 boards.
Last edited:
Your design will be more like a NAP100.
It's interesting that Naim don't seem to use groundplanes in their power amps. Also interesting that the NAP200 layout is much more spaced out than possible/usual.
Maybe the Chinese clones are not such a bad idea??
Julian Vereker said to me one day ( 1978 ? ). Many have copied the design, none sound like mine. You should have seen his face, it was like he won a gold medal in his mind. From being in various places in these forums I have noted how few really understand how the smallest details are critical. Within what is safe do it by ear. Car engines are the same. 99% of what an engine can be is determined in the casting process. 99 % of what it might be I doubt is ever realised. A friend was continuously tuning the old Mini engine. All he made really was Frankenstein versions, ugly in every way. What that said was the Cooper versions were the best possible version. Interestingly for insurance a special 850 CC version was allowed. That one was fun and cheap. 1100 head ( and it's larger valves ) skimmed 60 thou. Big carb as long as just one ( 1 3/4" SU from a Rover was an example ). Branch exhaust and special manifould. It can be done. That was perhaps going from 27 to 35 BHP inside the rules. What a sweet engine it was. Sorry to use a car analogy. That was 41 years ago. That engine and Micheline XAS tyres would still make me smile. 90 MPH and rally type grip. XAS didn't aquaplane unlike other fat tyres. It's all too easy to build a Frankenstein.
Ground planes are a good idea. Two sided PCB's were rare in the 1970's/80's. The Creek CAS4040 was hand drawn I am told. Remember to use heat relief pads if you are designing with one. Almost impossible to solder if not.
Ground planes are a good idea. Two sided PCB's were rare in the 1970's/80's. The Creek CAS4040 was hand drawn I am told. Remember to use heat relief pads if you are designing with one. Almost impossible to solder if not.
Last edited:
Again, looks like a quality kit from Caowei. I guess an issue here is that by the time the kit has been obtained, and a suitable transformer and chassis sourced and built then there won't be much difference in cost between the finished unit and a genuine 2nd-hand NAP 140.
For example, one here went for £150:
Naim Audio NAP 140 Power Amplifier | eBay
In this case I'd personally go for the genuine NAP 140 due to the minimal cost saving involved. For me the NAP 200 was an exception because of the price that even 2nd-hand examples command.
I guess it depends on the individuals' motivation for building their own units.
Chris
Last edited: