Watch the construction of 2 unit boxes. This is in the old-school Naim design format where the box is the heatsink and the construction would need to be at least 3mm aluminium, effectively in one piece for the base, rear and front or sides.
Hi Dave
Its been an interesting read this thread and I am thinking of implementing some of yours and Nigels suggested mods to the NCCs so would be good to have a chat with you some time, maybe the next NEBO event if your thinking of comming?
I would also very much like to hear your amp so I will PM you 🙂
Alan
I bottled out of the last NEBO, but could well go to the next one.
I would recommend two simple experiments:
1) Link out (short) the 22K resistor and have a listen
2) Link out the 100R emitter degeneration resistors and have a listen
I also tried the 0.22R output resistor instead of the LR. Not impressed - this seems to be where the Naim bass hump comes from.
Cheers,
Dave
I would recommend two simple experiments:
1) Link out (short) the 22K resistor and have a listen
2) Link out the 100R emitter degeneration resistors and have a listen
I also tried the 0.22R output resistor instead of the LR. Not impressed - this seems to be where the Naim bass hump comes from.
Cheers,
Dave
Last edited:
removing the two LTP degeneration resistors increases the open loop gain of the amplifier.
This needs a reassessment of the stability margins. If you are lucky the amp will simply ring on faster transients. If you're unlucky the amp could oscillate into destruction.
Ian I agree I am not a fan of bolting transistors just to a base plate, seen to many failure becase of this on PFM 😱
My plan is to buy a heat sink with vertical fins that will fit in or above the slot in the CB and mount the Sankens vertical.
If I cant get the heat sink to line up with the ventilation slots in the base and top plates I will drill rows of hole's to force a natural updraft ventilation like it should be 🙂
Alan
Perhaps you'll be just as brave trying that method. The clearance height remaining between the top of the PCB and the cover is a bit limited for standard sized heatsinks but assuming your case is nominally 410*300*88 mm, I imagine a heatsink of 250*70 with a good depth of fins would be fine provided there was free air inflow from beneath - though I doubt that is possible with this PCB unless fans are used.
A number of other black box designs from UK designers also use the case as a heatsink and some are fairly reliable but all are pretty marginal in the climate I live in, when the power is turned up. That means you can be down to a 10-20W amp and no summertime parties. I have a 1010 Exposure amp around here somewhere that uses only a 1 mm aluminium case for a sink but the secret there is to fit a 5mm plate as a heat spreader covering the area of the small PCB, the output devices and then some. A flat bar of 50*6 material about 200 long would be fine for the NAP 200 and not a lot different to all the early model NAPs, given the cases were also folded up from 3mm material.
Chinese boxes are often ≥ 3mm aluminium which is fine as this isn't a huge amplifier; just 2*75W. This case might suit you as it has 4mm sides and 3mm top and bottom in the style and finish of your preamp. The price isn't scary and offsets any freight cost, though I think the seller also has a European distribution warehouse which should save some cash: Aluminum Chassis DIY Case Enclosure Cabinet FOR Preamp DAC 430 95 330 MM | eBay
Fair point and I think I mentioned previously in this thread that I always check for stability after any mod. From what I can see the NCC200 and my 005CCN have plenty of stability margin and 3X loop gain only makes a marginal difference to the reactive load transient response.
BUT, do any mods at your own risk!
I think Nigels mods cover the stability issues after removal of the LTP Degen resistors.
Ian
The case will be either 2U in this case and the heat sink will be substantial being prown to overkill 😀
Alan
Ian
The case will be either 2U in this case and the heat sink will be substantial being prown to overkill 😀
Alan
hi Ian and all member,
Now I focus on Ncc200.I use 24-0-24v transformer 1.75amp.And it little hot. But still can use. After this I will upgrade to 30-0-30 transformer.
1.. I use wood box(diy), Apex preamp. And additional pot for volume control. Problem is. I hear noise some times have some times gone. Sound like "nyiiiiiiiiiiii". How to overcome this problem?
Action I take.....
1. Loop grounding-nothing effect
2. Unmounts bass and treble pot from wood body-reduce around 50%.
Need support from other member.
Now I focus on Ncc200.I use 24-0-24v transformer 1.75amp.And it little hot. But still can use. After this I will upgrade to 30-0-30 transformer.
1.. I use wood box(diy), Apex preamp. And additional pot for volume control. Problem is. I hear noise some times have some times gone. Sound like "nyiiiiiiiiiiii". How to overcome this problem?
Action I take.....
1. Loop grounding-nothing effect
2. Unmounts bass and treble pot from wood body-reduce around 50%.
Need support from other member.
Amplifiers generally need shielding (a metal case) and careful layout of the circuit board and wiring to maintain stability and low noise. A wooden box may be cheap but is likely to mean problems for the amplifier, like instability and being prone to electromagnetic interference (EMI). You can check this by moving the amplifier to another location and retesting.
A metal case is even more necessary for mounting the pots in the preamplifier. This connects the metal housing and shaft to protective earth (chassis) and shields against EMI. However, if you don't have a metal chassis, the correct earth point for pot. housings is important and should normally be at the main power supply chassis connection or in this case, star ground. Obviously, miniature plastic pots are not suitable for this.
The PCB is an important part of the circuit too and you should at least use one of the Ebay clone boards so that you have a reference layout. Use the specified semis as using the wrong types and particularly fakes, can also result in instability and poor sound quality. The VAS transistors are critical for this. Don't forget that the location of the output coil in the NCC200 version is important and you need to keep it as far as possible from power leads and traces around the output stage and particularly the input stage.
One of the best sites to read the general issues of amplifiers and audio electronics construction is the articles here: DIY Audio Articles
It will take a long time to wade through the info. there - enjoy!
A metal case is even more necessary for mounting the pots in the preamplifier. This connects the metal housing and shaft to protective earth (chassis) and shields against EMI. However, if you don't have a metal chassis, the correct earth point for pot. housings is important and should normally be at the main power supply chassis connection or in this case, star ground. Obviously, miniature plastic pots are not suitable for this.
The PCB is an important part of the circuit too and you should at least use one of the Ebay clone boards so that you have a reference layout. Use the specified semis as using the wrong types and particularly fakes, can also result in instability and poor sound quality. The VAS transistors are critical for this. Don't forget that the location of the output coil in the NCC200 version is important and you need to keep it as far as possible from power leads and traces around the output stage and particularly the input stage.
One of the best sites to read the general issues of amplifiers and audio electronics construction is the articles here: DIY Audio Articles
It will take a long time to wade through the info. there - enjoy!
24+24Vac @ 1.75Aac gives an 84VA rating.
That is small for a power amplifier.
That is TOO small for a stereo amplifier !!!!
use one times to two times the TOTAL maximum power output.
Two channels of 40W into 4ohms needs at least 80VA and it is ecomonmic to use upto 160VA.
BUT, I recommened that for low regulation you always adopt a minimum of 160VA for a power amplifier transformer.
If you built a 60W+60W into 8+8ohms amplifier, then use a 120VA to 240VA, but choose no lower than 160VA.
If you want this same amplifier to drive 4ohms speakers then the amplifier rating for 4ohms would be around 105W and that would require 210VA to 420VA.
Personally I would go for the 30-0-30V transformer and 300va minimum, when I build mine it will be 500VA
Just make sure all components are rated for over 45V.
If you are getting a screeching sound it could well be a faulty signal ground from somewhere, be carefull this can take out tweeters and send the amp into oscillation.
Alan
Last edited:
@sasek,
If I understand you correctly the noise issue comes and goes and you have increased heat.
Looking at the first of these if you are driving the amplifier continuously at high output levels, this will cause heating of the transformer which will cause mechanical noise. Good you have decided to change that it might kill two birds with one stone.
It is not clear if you fitted the additional output devices as indicated earlier. This practice was investigated by Naim long ago and dismissed - there being four lots of current gain and base-to emitter matches to be achieved rather than two.
You will need to exclude the possibility of thermal drift of the output stage as this is another possible cause of transformer loading
If I understand you correctly the noise issue comes and goes and you have increased heat.
Looking at the first of these if you are driving the amplifier continuously at high output levels, this will cause heating of the transformer which will cause mechanical noise. Good you have decided to change that it might kill two birds with one stone.
It is not clear if you fitted the additional output devices as indicated earlier. This practice was investigated by Naim long ago and dismissed - there being four lots of current gain and base-to emitter matches to be achieved rather than two.
You will need to exclude the possibility of thermal drift of the output stage as this is another possible cause of transformer loading
One way to assess whether the noise fault lies with the overheating transformer would simply be to turn down the volume. If that doesn't affect the relative level, period etc. of the noise, then its an another reason. You might also tell us what bias level (mV across the 0R22 emitter resistors) you measure.
I was doing some thinking about the NAP VAS.
People might find some more straight forward information in this regard at http://www.talkingelectronics.com/p...ctical_electronics_handbook_sixth_edition.pdf starting at page 132. The book is authored by Ian R Sinclair and John Dunton.
As far as output stages, go Linsley-Hood looked at several "Gain Stage Designs" in his book "Valve and Transistor Audio Amplifiers - reprinted in 2000 page 182 et seq. In all of these cases the usual constant current source loading the last device in the voltage amplification stage is replaced by a 2K7 resistor - a deliberate choice to emulate the load of a push-pull Darlington output stage.
As far as transistor gm is concerned both authors consider a gm for a small signal transistor figure of 40 mS (mA/V) to be practical. According to Hood the range varies from 25 mS (probably low current gain power devices) to 40 mS.
Last edited:
I looked at the 40mS rule of thumb with calculations between 1 and 10 m.a. where this result is independent of transistor Beta. For this the authors have preferred a thermal voltage of 25 mV - possible junction temperature - changing this to the alternative 26 mV room temperature value makes little difference (25/26 times 40 = 38.46).
Below 1 m.a. the mS figure is in proportion so 0.5 m.a. gives 20 mS. Apparently Hood drew a lower current line of 0.62 m.a. for his 25 mS lower limit - circumstances relevant to voltage amplification rather than the use of power transistors as current amplifiers.
In that regard things begin to ascend from 10 m.a.
Below 1 m.a. the mS figure is in proportion so 0.5 m.a. gives 20 mS. Apparently Hood drew a lower current line of 0.62 m.a. for his 25 mS lower limit - circumstances relevant to voltage amplification rather than the use of power transistors as current amplifiers.
In that regard things begin to ascend from 10 m.a.
This is incomplete - there is a jump in the operating current from the LTP transistors and the Vas.
10 m.a. is the operating level for the later where gm= 400 mS - the rule of thumb being gm (mS) = bias current (m.a.) times 40.
I would like to add another rule of thumb. If driving a highly reactive load double your voltage limits ( 230 V is wise ). I had practical experiance of this two days ago. Driving some TIP2955/3055 at 76V ( +/- 40 V max ) into a transformer the output stage caught fire on switch off ( the transfomer takes the devices above the voltage rail due to stored energy, protection diodes were fitted ). This wasn't a surprise and had given me the data to continue. Totally stable and noise free to 16 Mhz. The VAS is MPSA 42 which is fine as are the 2 x BC327-40 input stage ( only 45V and it's enough ). This is an opposite layout to a Naim in that the input stage is PNP. The idea is a future proof design. The MJE15032/33 can be MJE 340/350 if fitted facing the other way. For this application they would be fne.
The proposed real version is. The gain of the output devices means that 2N5551G should be OK. 6 mA should be enough VAS current. I am assuming gain > 50 for the outputs and a realistic compound gain of 1000. I might need 2 amps @ 18 Vrms.
Input 2 x BC556B ( 65V PNP )
VAS 2N5551G ( 160 V NPN )
VAS CCS = Bootstrap
Drivers
MJE15032/33 250 V u= > 70
Outputs
TTC5200 ( N ) 250V
TTA1943 ( P ) 250 V
2 x 0R33 5W.
This is a very simple amp to drive a motor at 30 watts continuously ( that's like 150 watts of music ). It was depressing to find how few transistors there are now. The output devices have the greater choice. I will fit a Vbe bias control as the amp could be used for hi fi. For motor drive 1 x LED seems OK. Although it doesn't track the output it holds it down far enough at all times. With feedback added it is enough at 100 Hz to give OK distortion.
I also tried T03P MOSFET's ( not audio types , like IRFP240 in P and N ). They didn't like the transformer.
The proposed real version is. The gain of the output devices means that 2N5551G should be OK. 6 mA should be enough VAS current. I am assuming gain > 50 for the outputs and a realistic compound gain of 1000. I might need 2 amps @ 18 Vrms.
Input 2 x BC556B ( 65V PNP )
VAS 2N5551G ( 160 V NPN )
VAS CCS = Bootstrap
Drivers
MJE15032/33 250 V u= > 70
Outputs
TTC5200 ( N ) 250V
TTA1943 ( P ) 250 V
2 x 0R33 5W.
This is a very simple amp to drive a motor at 30 watts continuously ( that's like 150 watts of music ). It was depressing to find how few transistors there are now. The output devices have the greater choice. I will fit a Vbe bias control as the amp could be used for hi fi. For motor drive 1 x LED seems OK. Although it doesn't track the output it holds it down far enough at all times. With feedback added it is enough at 100 Hz to give OK distortion.
I also tried T03P MOSFET's ( not audio types , like IRFP240 in P and N ). They didn't like the transformer.
I don't want to take this too far away from the main topic, so is your illustration in reference to an audio application such as an amplifier driving a gramophone motor via a transformer.
It was to tell people to consider a reactive load as something that requires a surprisingly high voltage spec. A loudspeaker is an excellent example of a reactive load. Up to the VAS maybe not so important. The more interesting fact is that good quality easy to find transistors are becoming harder to find. I was rather surprised how good the 2N5551G is. It was in the past $3/100, therefore we tended to think it was cheap and low grade. BC556B is a quality device. The other highly important thing is the voltages I was using are like the Naim. My proposed design will go to +/- 63 V, I shall try it.