555PS DR Towering Inferno – Resistor Values Needed Please

[Cloth Ears]

Well luckily the whole Twin Towers Naim rack was not quite a blazing inferno but I have discovered a seriously burnt out section of circuit board within the 555PS DR feeding my NDS which is very alarming. This is the circuit responsible for discharging the pair of 15,000 uF capacitors feeding the split voltage analogue supply regulators which has seriously overheated and burnt the PCB, capacitor discharge resistors and controlling transistors.

I discovered this damage quite by chance after I was inspired by some old posts from 2018, 2019 and 2020 on the advantages of adding a second PSU on the NDS and ND555 streamers. The second power supply feeds the analogue and separates it from the digital supply. There was discussion around using Teddy Pardo or Naim PSUs and various Burndys.

I opened up the 555PS DR to check out the wiring to the second Burndy socket with a view to making a clone of the 555PS containing just a transformer, rectifier board, 2 capacitors, 2 regulators and a single Burndy socket which looks quite doable.

When I inspected the rectifier board I noticed how brown it was around the busy section nearest to where the DC wiring leaves the board. I removed the PCB to have a look underneath and was horrified to see the resistors and transistors well fried. The amazing part is the 555 PS and NDS work just fine and sounds great. I have measured all the supplies from the transformer right through to the Burndy output pins and everything is running OK.

The Rectifier board is quite simple and comprises of 7 pairs of diodes feeding 2 onboard smoothing capacitors and 5 screw terminal caps off board. I gave it a good dose of reverse engineering and drew out the circuit which has given up the magic smoke. After cleaning it up a bit I can read the part numbers on the 4 transistors and identify most of the other parts but the 7 resistors in the photo below have lost their coding rings. Can anyone with a 555PS help identify their value please. You will need to remove the 5 small torx screws before turning it up from the rear of the chassis. These are the 7 resistors which I need the values for in the picture below.

Prior to cleaning up the damaged area I tried to measure the 7 resistors but as expected got very variable results. The 7 damaged ones are a group of 3 low value in parallel, a single high value, then another group of 3 low value in parallel. Discounting the scorched 47K resistor they measured 22R across the left group of 3, then 2 - 3M for the one in the center and 22R across the right group of 3. This would put each resistor in the two groups of 3 at 68R each. The following picture is the discharge circuit as I have drawn it out and this is how I think it works.

At the left side a 22V AC feed from the same windings supplying the rectifiers and 15,000 uF capacitors for the analogue circuit is rectified and trimmed to -1V by the 4 small diodes. This negative voltage is applied to the base of TR2 which forms a Darlington pair with TR3. The collector of TR3 holds the bases of TR1 and TR4 so they are biased off and the capacitors run at their full voltage of + and – 30V.

When the power supply is turned off the voltage at the four diodes is no longer negative and is pulled up to + 1V by the high value 2 – 3M resistor. This stops TR3 collector from holding the bases of TR1 and TR4 down and they are pulled up by the 47K resistor from the +30V rail. This causes TR1 and TR4 to switch the 2 groups of 68R resistors between the +30V and -30V rails discharging the two 15,000 uF capacitors. Obviously the transistors are now toast so the circuit cannot function correctly but I can still measure the voltage at the base of TR2 flip from -1V to +1V at the moment of switch off.

There is an almost identical three transistor circuit in my olive Supercap with fewer components and just one transistor connecting two 470R resistors in parallel across the single 10,000 uF capacitor which feeds the 12V / 24V regulators for the control supplies. The Supercap discharge circuit is around 22 years old and has been on and off more times than a whores draws without any problems.

In the 555PS I am assuming at least one of the transistors has gone faulty at some point and allowed the bases of TR1 and TR2 to be pulled high switching the resistors across the supply whilst the 555PS is still on causing a major overheat and luckily it did not set the house on fire.

I am particularly keen to know the exact value of the two groups of discharge resistors because this will tell me where Naim have pitched this circuit. From my measurements of the blackened resistors they appear to be three 68R in parallel which means TR1 and TR4 are each switching 60V across 22R7 at the moment of power off. This will cause the tiny transistors to carry 2.67 A each at the moment of initial discharge. The data sheet for these ZTX753 shows absolute max of 2A but also states up to 6A pulse. From the resistor point of view if these are 68R each and I’m driving my Abacus correctly they will be asked to pass 0.8823 A at the moment of switch off and dissipate a momentary 53 Watts each. No wonder they complained.

Can anyone shed any light as to why this circuit exists. I can appreciate that in the case of the S/Cap there would be a desire to drop out the control/mute relays etc. as the voltage rails fall but cannot see how this fits in with the NDS /ND555 analogue circuits unless it’s a legacy from the CD 555 which this supply was originally designed for.

Hope you can help.

Thanks.

 

[hp1]

I think the circuit is to discharge the caps quickly to prevent any unprofessional noises and unwanted voltages at switch off.
I did a mock up ps for the output filters on nd55 and cd555 which did not have this device and could here no nasties at switch off
on either player.

I think the 3 resistors in parallel are 620 not 68 ohm as mine measure 207 ohms for the trio.

 

[Cloth Ears]

Yes I would agree it could be there to remove possible unwanted noises at switch off although I always power off the main amps first so don't really hear anything. One other reason I considered is this supply is a split polarity, i.e +30V and - 30V so it may be to ensure the voltage is reduced to near zero as soon as possible to prevent the situation where one half of the supply is decaying faster than the other which could produce all manner of strange effects in the signal circuits and might be detremental to the well being of the components.

Thank you very much for measuring your 555 PS circuit. That's really valuable information and it's quite reasuring to find that Naim are not trying to discharge the caps nearly as fast as the screwdriver across the terminals method. I knew that measuring fried resistors would not be reliable and your measurements are much more in line with what I would expect.

Do you have a reading for the high value resistor that is situated between the two groups of three discharge resistors?. I think it is about 3 megohm and although mine looks well fried I think that is only colateral damage as it should not have received any fault current if my understanding of the circuit is correct.

I will run the numbers again based on your measurements and see how that sits with the ratings of the transistors and resistors but it should be a whole lot better. I will get hold of some new components and make up the same circuit for offline testing - I don't think the original board will take another fireworks display.

Thanks again for the information.

 

[hp1]

I will lift the pcb and report back.

 

[hp1]

H

 

[hp1]

Hi,
Its 2x 3 x620 ,100k,47k,1m,3m9.

 

[Cloth Ears]

Hi hp1,
A really big thank you for taking the trouble to find out the resistor values.

That all makes perfect sense and will be of enormous help in rebuildIng the circuit. The high value 3M9 is exactly the same as the S/Cap discharge circuit.

On the actual discharge resistors at 620R each I wonder if mIne started life as 62R then after getting well overheated became 68R i.e. a production error at Naim.

The thinking behind that is my 555PS was made in 2012 and I bought it ex dem fom my dealer in 2014. When I first got it home I could not resist taking a peek inside and I remember seeing a slight browning around the PCB soldered joints on the discharge resistors. At the time I didn't know what the circuit was and I just assumed it was a bit of excess brown flux. I took a pile of photos at the time which should still be on my NAS. I will hunt them out later and compare them with what I can see now.

Once again a big thank you.

 

[hp1]

No problem at all.

 

[Cloth Ears]

Well I checked the photos on my NAS and here is a picture of the PCB taken shortly after I bought the NDS and 555PS in 2014. What I remembered as slightly brown marks that I thought was flux were in fact blackening around the discharge resistor soldered connections. How did I miss that!

Here is a similar picture taken a few days ago with a better camera which clearly shows the same blackening around the connections so it appears that the circuit had already gone faulty before I received the PSU.

That clears up the big mystery of how it went faulty without me knowing. All the HiFi is on 24/7 except the usual pending thunderstorms and away from the house for more than 24 hrs so I would have noticed the burning smell.

I will give the faulty circuit a detailed forensic examination as far as I can before I contact my dealer and Naim technical.

I intend to start by obtaining the parts and making up an identical experimental circuit that I can run through it’s paces measuring temperature rise and power dissipation in the resistors.

I am still intrigued by the discharge resistor value which hp1 has confirmed is 620R. My damaged ones are reading around 68R but surprisingly none are open circuit. Every magic smoke event I’ve seen, which are quite a few, always ends up with the resistors value going higher not lower but of course that depends on the construction of the resistor. I will also try a version of the circuit with 62R resistors to check out the unlikely possibility that incorrect resistors were fitted during manufacture.

I propose using the pale green Vishay SFR25 0.4W resistors which I believe are the same ones Naim use. I will try both values in normal and overload conditions and see how they change after overloading.

I am rather disappointed by the lack of safety engineering in the Naim circuit. Tiny 0.4W resistors fed by a 25A rectifier with 60V across two beefy 15,000uF caps propelled by a 500VA transformer is asking for trouble if it goes wrong.

I am considering replacing the 6 off 620R 0.4 W resistors, which form two groups at 207R, with two single 210R high wattage jobbies. These would need to be capable of dissipating about 17 W each under fault conditions i.e. TR1 and TR4 biased on or gone Harry Shorters whilst the PSU was still running.

I was thinking of something like two Caddock T0-220 case power resistors mounted on a small aluminum upstand with a manual reset thermal trip at 60C using the same fixings on the other side to interrupt the -30V line to TR1 and TR4.

When the thermal switch operates there would be a voltage across it to generate a fault signal from. I could use that to operate a circuit to switch off or flash the Naim logo.

I will get some parts ordered - watch out for the flames.

 

[hp1]

In my opinion there is nothing wrong with the design of the circuit, I have 5 of them and they are always being switched off without problem.

 

[a.palfreyman]

620R across 60V is 0.1A at discharge (instantaneous) but will quickly fall as you've already said. P=I(squared)xR = 0.1x0.1x620 = 6.2W instantaneous. You should just about be able to get 2W resistors in there, which is a considerable improvement over the original 1/2W or thereabouts and you could stagger the heights on the board to allow them to 'breathe'. As these are just to discharge the caps, you could also increase the value a little. I'm sure 1k each (so 330R for the three) would suffice. You'd then need 0.06x0.06x1000 = 3.6W instantaneous and 2W rating would be OK

 

[hp1]

I dont think the 210r resistors are the problem here, much more likely a transistor which has gone closed circuit.

 

[Cloth Ears]

Sorry guys I’ve been offline for a while. I would agree with hp1 there is probably nothing wrong with the basic design of the circuit but only when everything is working correctly.

I do think Naim could have used higher wattage resistors that get less stressed as a.palfreyman's calculations shows.

I have yet to receive all the parts I need to build some test circuits but the first thing I would like to establish is what happens to the six 620R resistors when subjected to 60V until they catch fire. Does the resistance drop to around 68R each as mine appear to have done or was my 555PSU fitted with incorrect value resistors at manufacture.

I agree the most likely cause is a transistor failure that has left the six 620R resistors across the 60V. My concern is that Naim have not made any safety provision in this circuit. A failure in any single part of the ‘hold off’ circuit which keeps the bases of TR1 and TR4 biased off will result in the circuit activating during normal unattended running.

A couple of days ago I was trawling through all the Naim pics I could find looking at their discharge circuits when I came across the picture below of a NAP 500 PSU. I am only guessing but it looks very similar to my plan of using high wattage resistors and a thermal cutout to disconnect the discharge circuit if it fails. By it’s location near the mains wiring it looks like Naim are turning the whole PSU off. MMmmm

The Plot Thickens

 

[Cloth Ears]

I have had another trawl on the net and found a pic of a NAP 500PS which clearly shows the discharge resistor is a WR10 15 Watt 220R. It looks like the thermal switch simply interrupts the incoming mains from the IEC socket to the mains switch.

The thermal sensor looks similar to the type fitted on the NAP250 and could be a self resetting one. If that's true, and the transistors in this circuit fail, then the PSU will just cycle on and off as the resistor heats up and cools down.

For my experimental circuit, which I intent to test under fault conditions, I have ordered a 45deg C heater style switch with manual reset. I was considering just interrupting the -30 V line but the switch is mains rated so I may interrupt the mains instead.

It just depends if it has an effect on the sound quality. However, burning the house down may be more detrimental to the sound

 

[hp1]

Naim has always been a belt and braces sort of company and again I cant see anything wrong with the 500ps except for the inrush current across the power switch which is fine as long as the switch contacts are in good condition .

 

[Cloth Ears]

Hi hp1,
I'm not sure if you mean the manual on/off switch or the thermal switch fitted to the resistor. I would expect they are both fully rated for the inrush.

My point was that should the transistor circuit fail in the 500PS discharge circuit then the resistors will overheat and operate the thermal cut-out switch. If like I suspect it's an auto reset cut-out it could be cycling on and off for some time until Its noticed. A manual reset cut-out will of course keep the PSU off until its reset.

The cycling should not bother the mains on/off switch as it will stay closed and not be making onto the inrush current - just passing it through each time.

 

[hp1]

Hi, Mr Ears,
I was talking about the main power switch as they do fail if not maintained .. I have looked at the high powered resistor and what it does and think it is there to protect the transformer . If there is too much continuous current from the transformer the power resistor heats up and sets off the thermal trip below, again Naim being belt and braces.

 

[hp1]

Perhaps the power resistor softens the transformer inrush to save the power switch also.

 

[Cloth Ears]

Unfortunately I don't have a 500PS and I'm just going by the picture but it looks like the wiring goes directly to the capacitor board near the third set of caps from the back. Its definably a 220R resistor so I still think it must be to do with the discharge circuit as that would be a very high value to have in circuit with the transformer primary.

 

[hp1]

Its been a long time since I looked at one of my 500 ps s so you could well be correct.