Cambridge Audio A3i repairs and mods

I've designed that amp more than 12 years ago. As it happened, it was designed to be a 40W per channel amplifier but at the last moment it was promoted by the management into 60W just by boosting the supply voltage. That and also poor quality PCB and some components contributed to a common fault - as a result of a local board overheating one of the LED's (usually LED2) or Q8 transistor next to it (BC640) goes open circuit. Thankfully, protection circuit in this case just disables the output. It is useful, if you came across this amplifier, replace both LED2 and Q8, mounting them at least 10-12 mm from the PCB.

I've bought an old and broken A3i on ebay couple of months ago - to see if I can modify it to improve the performance. After fixing the usual fault (it was Q8 this time) I found that few simple mods could considerably improve the amp performance.

To get a baseline, I've measured the amp in it's original condition (except the transistor replacement). A3i gave a very typical set of measurements:

60 W into 8Ohm

at 50 W into 8 Ohm THD was:

0.003-0.005% upto about 5 kHz
0.01% at 10 kHz
0.022% at 20 kHz

However, as the first picture (1kHz 20V RMS 8Ohm) shows, there was a considerable amount of interference from the rectifier.

Today I did a few mods on the old A3i - replaced rectifier diodes with Shottky and replaced D210 and D211 (diodes between VAS supply and O/S supply) by 10 Ohm 1W resistors. As a result, most of the interference have disappeared and more importantly, the THD measurements improved significantly:

below 2 kHz the distortion was at the limits of my ability to measure them. My sound card ADC has got residual distortion about 0.0003% for -2 dB input (equivalent to 20 V RMS on the graphs), my generator is about 0.00015% at midrange. The amp measured consistently near 0.0005% for 2kHz and below, 0.002% for 7 kHz, 0.007% for 10 kHz and 0.02% for 20 kHz.

Providing that the gain of the power amplifier in A3i is quite high - 57 (about 35 dB) the performance is good. Most of the distortion at HF is due to a crossover point - result of a simple two-stage output follower. Using a triple follower could reduce the distortion further. However the VAS has excellent linearity for such a simple circuit.

Alex

P.S. I attach the schematics of A3i power amp section and two graphs - on the left is for the unmodified amp (THD = 0.002%) and on the right is after modifications (THD = 0.0005%).
 

Attachments

  • a3i_modification.gif
    a3i_modification.gif
    95.1 KB · Views: 6,703
Disabled Account
Joined 2007
Hi Alex,
This is a very good performance improvement! Nice looking schematic too.
I have been looking for a good medium power amp for the mids in an active 3-way and a place to use some vintage outputs I harvested from an old Pioneer amp.
Do you think the 2SB705A/2SD745A would be a good substitute for the 2SA1216/2SC2922?
 
MJL21193 said:
Hi Alex,
This is a very good performance improvement! Nice looking schematic too.

Thank you!

MJL21193 said:
I have been looking for a good medium power amp for the mids in an active 3-way and a place to use some vintage outputs I harvested from an old Pioneer amp.
Do you think the 2SB705A/2SD745A would be a good substitute for the 2SA1216/2SC2922?

These are smaller than 1216/2922 and can be used with a lower supply voltage, say +/- 30-32 V. It will give you about 35-40W into 8 Ohm

Alex
 
Hi Alex,
I have an A3i that I want to keep because I really like the sound - congratulations on your design! Everything was perfect until a week ago but now when I turn the amp on or off its gives a very disconcerting thump through the speakers. Still works great while it is on. The first few days I got around this by pressing "mute" before turning on or off but now the mute also does not work! I want to keep using the amp (I really like it) but am a bit nervous for my speakers (old KEF Reference 105's of which I am also very fond). I have searched the forums and downloaded already the schematic for the protection and power supply, as well as your link for a recommended relay replacement. I am asking for suggestions where I might start to look for the problem - the relay might be the first? I am not a technician but have friends at work who certainly are. I just need a suggestion where to start before I approach them for help. I note from previous threads that you were prepared to send the A3i schematics as an attachment - my email is mancktelow@swissonline.ch
thanks,
Neil
 
nman said:
Hi Alex,
I have an A3i that I want to keep because I really like the sound - congratulations on your design! Everything was perfect until a week ago but now when I turn the amp on or off its gives a very disconcerting thump through the speakers. Still works great while it is on. The first few days I got around this by pressing "mute" before turning on or off but now the mute also does not work! I want to keep using the amp (I really like it) but am a bit nervous for my speakers (old KEF Reference 105's of which I am also very fond). I have searched the forums and downloaded already the schematic for the protection and power supply, as well as your link for a recommended relay replacement. I am asking for suggestions where I might start to look for the problem - the relay might be the first? I am not a technician but have friends at work who certainly are. I just need a suggestion where to start before I approach them for help. I note from previous threads that you were prepared to send the A3i schematics as an attachment - my email is mancktelow@swissonline.ch
thanks,
Neil

Hi Neil, from what you've describe it does sound like a relay or the control circuit is faulty. Normally there is a delay for the relay switch-on so you won't hear the thump on power on or off. Now you're saying that even the mute button does not switch the sound off? In this case the relay is probably constantly engaged - stuck in ON position . I've sent the schematics - hope that would help your friend to fix it. If you need more help with the amp - just email me.

Cheers

Alex
 
Hi Alex and Alls,

first of all, congratulations for the project.
I think if you want changing the output BJTs you should change also the drivers. You could try other more performing BJTs than BD139/140.
I think you could change(for example) the BD139 and BD140 with MJE15030/31 but for this modification you should increase the 68pF capacitors for more phase margin (like you know). Moreover R21 could be changed with a 68 ohm for increase MJE15030/31's quiescent current.
I have seen that fm acoustic amplifiers have big drivers.
I have tried this modification in power amplifiers (that I have designed) with success.


PPA
 
ppa said:
Hi Alex and Alls,

You could try other more performing BJTs than BD139/140 .

Hi PPA,

the original Philips BD139/BD140 are very good devices. However what is now sold under these numbers is not good at all (and don't have any relation to the originals) . So now I would not recommend using BD139/140 if someone would like to build this amp, and replace them with other modern driver transistors, MJE15030/31 or similar. I will try to experiment with drivers replacement as well as with other possible small mods when I'll have time.

Alex
 
This circuit is a very interesting one. Very unique VAS :)

x-pro, I presume this amp was designed to a budget. If you could make more exotic device substitutes and tweaks, what would you have done? My guess is you would have used a triple stage as you have already mentioned, and better driver transistors.
 
jaycee said:
This circuit is a very interesting one. Very unique VAS :)

Thank you, the VAS is indeed quite interesting - it combines a very high open loop gain and a good linearity for such a simple circuit. I've used a somewhat modified version of it (with fully balanced cascoded JFET inputs) in Creek Audio A52 series of power amplifiers.

jaycee said:
x-pro, I presume this amp was designed to a budget. If you could make more exotic device substitutes and tweaks, what would you have done? My guess is you would have used a triple stage as you have already mentioned, and better driver transistors.

These are obvious ways of potential improvement. A3i was made to a very tight budget. In general, this VAS could be easily coupled to another good quality output stage with good results.

Alex
 
ppa said:
Hi All,

Alex, what do you think about the change of bd139 and bd140 with the MJE243/53 ? their pin sockets are the same but power and curret are both bigger in the seconds, moreover MJE243/53 are made by ONsemi and this is a guarantee.

Pier Paolo

Hi Pier,

yes, these look like a suitable replacement, thought according to the datasheet there is a noticable difference in a current gain between n-p-n and p-n-p devices. I may order some from Farnell to try and measure.

Cheers

Alex
 
ppa said:
Hi All,

Alex, what do you think about the change of bd139 and bd140 with the MJE243/53 ? their pin sockets are the same but power and curret are both bigger in the seconds, moreover MJE243/53 are made by ONsemi and this is a guarantee.

Pier Paolo

Hi Pier!

here is an update. I've received some MJE243/253 and tried these in A3i circuit. They work very well - certainly better than modern versions of BD139/BD140. I also have measured the gain of these devices and from 50 of each all PNP have gain between 140 and 150, and NPN are mostly at 130, with a range from 100 to 150. I plan to make a redesign of A3i circuit for DIY and will probably use ON Semi drivers and output devices.

Alex