Cambridge Audio A3i repairs and mods

Hi all, my name is Sebastian. I am from Germany.
I have just purchased a broken A3i to fix it, as I have found this here thread.
On my first inspection I came across LED102 not working but thats not my first spot of concern.
I quickly realized that the heatsink was heating up from the front (I suspect Q114 as the culprit).
The amp is drawing about 150W from the socket without any load, even with the volume knob turned all the way down.
Any ideas on what to do are greatly appreciated.

Thanks in advance,
Sebastian
 
With the LED off the current source for the VAS is not working properly. The output stage may well be oscillating, explaining the heat and the power draw.

Test both of the LEDs (or just replace them) and their associated BC639/BC640 transistors as well. The original poster of this thread, x-pro, designed the amplifier so he knows what he is talking about 🙂
 
I've finally got round to looking at the A3i and someone had put Q5 and 6 in back to front for some reason, on both channels! There's clear markings on the board. They got fed up and sold it on. Replaced them and the other front end ones, and burnt out emitter resistors on one channel. Works fine now. Next job is the upgrades recommended by x-pro. It's fairly easy to repair this amp and well worth the effort.

I may be being fussy but the gain is a bit low compared to other amps of this vintage. Was it designed this way for high output CD players?
 
I have on my bench an A3i which crackles, mainly in the LH Channel. When the volume control and switches are operated, loud crackles are also heard. The amp also cuts out (protection circuits operate) for a few seconds then re-sets as the controls are operated. I have applied Servisol 10 to the pots. & controls, but the fault still persists. The supply rails measure +39/0/-39v. with my DMM, and there is a DC offset of about 60-70mv on the LH channel speaker terminals, more than that on the right channel o/p. I haven'y yet carried out any further tests. Would I be correct in assuming that the LH o/p transistors, and possibly others in the Left Channel, are breaking down, and would any of the mods suggested by Alex (x-pro) be advisable as a cure or improvement ?
 
Thanks for the suggestion, jaycee. C8/C108 according to my copy of the circuit are 100uf/35v electrolytics on the base of Q2/Q102, which, if leaky or intermittent could indeed cause this kind of fault, particularly if they are Tantalum Bead types - I don't have the amp. in front of me at the moment, but will check later, and post further when i've done so.
 
A further comment/question. In a thread posted some years ago by x-pro (Alex),who designed the A3i,he said that he'd replaced the original rectifier diodes (Type PW04) with Schiottky Rectifier diodes, but he didn't specify what types. Would that be advisable, and what types should I use?
 
In a thread posted some years ago by x-pro
It is this very thread. 🙂

It's easy to separate the pre and power amps, just unplug the link at the front of the power board, which will help tracking down the problem. I doubt it's the o/p transistors. The protection circuit will operate if there is a DC offset on the o/p. C8 is the AC grounding capacitor in the feedback path, so if this is leaky it will upset the DC offset. It's easier to just change components rather than test them, at least when they're so cheap.

The Schottky rectifier diodes improve noise isolation I think, so are for performance improvement rather than reliability. Sorry but I can't recommend any. The other upgrades are more important and I'd suggest increasing the value of R16 and R18 to 3K9 ohm as well as increasing their
power handling.
 
I've already dismantled the amp. and removed the PSU/Power amp. board, replaced C1 & C8, (C1, which had previously been replaced, had been fitted the wrong way round, so may have been the cause of the fault, but as the PCB is fiddly to remove & refit, I'll most likely change the corresponding components (C101 & C108) in the other channel while it's out. With your post # 243 in mind, spandrel, I'll check that the transistors Q5 & 6 are fitted the correct way round. With care, I can apply power, and connect speakers or dummy loads with the amp. PCB inverted, and test it before refitting & reassembling the A3i.
 
Thanx for sharing. Interesting that HK did a similar VAS ~40 years ago but seems it went largely unnoticed. I'm not sure there is a net advantage. I see a lot of clever topology that fails to provide real improvements.
I would not use LED voltage references. Seems like this amp will clip at least 4 Volts from the rails. Who cares? Well, I do. Maybe if you use a boosted VAS power supply.
I would add a ~47uF cap across R21, but if you are using fast power transistors that may not help much.
I would add a small resistor in series with C7 to provide a bit of feedback help with cross-over distortion.
I'm painfully aware that comments based on theory are dangerous, but some times the problems with implementation are avoidable, usually by placing the right resistor in certain places.
Thanx again for sharing.
Steve
 
The Audiolab 8000 and early Arcam used the same arrangement of dropping the voltage to the VAS stage, but NAD used a higher voltage in the 3020. Early Cambridge P series amps didn't do either.

x-pro's suggestion of an 80mA quiescent current might be to reduce crossover distortion. No problem for the o/p devices.
 
I'll bear that in mind, spandrel, although the amp. does work, but the protection circuits cut in as the volume control is rotated, then re-set a minute or so later. The 4A fuses have, so far, survived, but no point in blowing them during testing. I wouldn't say that someone else had 'made a mess of it', but whoever previously replaced C1 & C101 did managed to fit the former the wrong way round. This may have been the cause of the fault, of course, as the cap. would break down due to reversed polarity. Not having seen the underside of an A31 PCB before, I'm also wondering if the two lengths of insulated wire connected between the relay contacts and the output stage center rails have also been added for some reason, as there appear to be factory fitted bare wire links on the PCB to make these connections already. (The relay works as intended, b.t.w.-it hasn't been bypassed) The D.C. offset on the L.H. o'p is higher than that on the R.H. side, but is < 60mV. When I've solved the fault, I'll obviously re-check things, and, if necessary. set the bias as per the makers instructions. I did send a PM to x-pro (Alex), the designer of this amp. a few days ago, but have had no reply yet. Maybe he doesn't visit the forum anymore!
 
If the wires to the mute switch are still in place I'd say the extra links were a belt and braces addition to lower the impedance of the connections. Quite unnecessary I'd say unless the original connections are dodgy. I wonder why there's a speaker mute switch?
 
I checked the connections earlier this afternoon, and, with the under-board links unsoldered, there is no connection between the relay contacts and the center rails (emitter resistors) of the o/p transistors, so the wires must be original. The component side uninsulated links connect to other parts of the circuit. This is why a PCB layout drawing would be useful, but none seem to be available. I also found that, because the protection circuits are on the (temporarily disconnected) control PCB, the relay doesn't operate. As I don't have a long jumper lead to connect the two boards together with the amp. board inverted, I'll either have to temporarily solder links in place or bypass the relay. The latter will be easier, and, from the voltage readings I took, should be O.K. for test purposes. As to why there's a speaker mute switch, I don't know either- I removed the mute switch wires to make it easier to invert the board. (oops! I see now why the relay doesn't operate-it's because the speaker mute lines are in series with the relay coil!!!) A further point - C1/C101 are rated at 63v on the diagram, but I only have 35v 10uF Caps. As I cannot see how there could possibly be anywhere near 63v across these caps., I'm inclined to fit 35v ones, unless anyone knows why I shouldn't?
 
Last edited:
Having replaced c1/c101 (one of which had been inserted the wrong way round!) and c8/c108, there is still a fault present as the amp runs very warm after only a few minutes and readings taken around the o/p pairs suggest a fault in the BD139/140 area as only about 250mV is present on the B-E junction of one o/p transistor, whilst the DC offset on the suspect channel is over 60mV, though it can be adjusted. Any suggestions as to the cause of this ( The amp. is running without the control & preamp PCBs connected, b.t.w.)
 
Umm... there's no adjustment for DC offset.
Check there's quiescent current, normally 24mV between emitters. The presets should be set clockwise to start with no current. There should be about 2.4V between the bases of the drivers. Check there's current flowing through the VAS transistors, Q8 and Q10. It's easiest to measure the voltage across the emitter resistor R17.

Now for my own sorry tale. I swapped Q8 and Q10 VAS transistors for MJE243/253. I got the pinout of the old BC630/640 from the Towers transistor book. Tried it out with lightbulbs and it was drawing too much current. Huh. Checked I'd fitted the new ones correctly.... what's the problem? Tested the actual old transistor pinouts. GAAAHH! The book is wrong. Got the correct pinout from the datasheet and even traced the circuit tracks to be doubly sure. Fitted new trannies and it still doesn't work. Something else has blown. Grrr.

So yeah, I've packed it away for another day. Oh well.