Folks, I want to share my journey with Quad 306, in case it might help someone else in the future. Firstly, I am in no way an electronic expert but I have been tinkering around the amplifiers for about 30 years, so I have learnt a few basics along the way.
I acquired a Quad 306 (Revision 2 PCB) from a friend. He was about to dump it in the recycling because one channel was dead and the fact that he had bought a new Marantz amp. It has since been lying around with me for a couple of years. The past month I got an itch to feel the smell of the solder again, so I opened it up to see the internals. I found a few burnt resistors and after doing some research, I replaced the resistors and turned on the amp again. I was amazed to find that's all that was wrong with the other channel, which was now working with no issues. I was also very impressed with the sound quality coming out of it. This made me think if I could improve it further.
The first step was to figure out why the resistors got burnt, but the answer was right in front of me in the Quad 306 service manual. Page 10, note (C) - "The centre-tapped DC Line is floating and some faults may cause this to shift putting up to 80v on one side possibly damaging R30/R31 (2K2ohm)." However, I also had R22 (560R) smoked out, not sure why.
So after replacing every single resistor with better-specked Vishay resistors and doing some upgrades, I am gobsmacked with how good this thing sounds. I am literally in awe and just in love with it. Here is the list of things I have done;
I noticed, since putting Schottky diodes across the main capacitors the amp doesn't heat up much at all. It has most certainly brought the working temperatures down a lot. At low volumes, it doesn't even get warm and stays cold to the touch. I can have the amp running for hours and it barely gets warm. Whereas before it use to get quite a bit warm/hot to the touch. The difference in heat reduction is huge.
Few warnings for anyone who might be working on these little gems
All in all, I am still in shock by how good this amp sounds, I am hearing instruments in the songs, which I didn't even know existed before. Now I am wondering if my crappy speakers sound so good with it, how would it sound with decent quality speakers. Looks like it's about time to DIY a 2-way speaker system. This is it, folks, I thought I might share my results just in case it might help anyone else here. Happy DIYing and thanks to you all who contribute to the forums, it helps the likes of me (Civil Engineer by profession but with a DIY audio bug).
I acquired a Quad 306 (Revision 2 PCB) from a friend. He was about to dump it in the recycling because one channel was dead and the fact that he had bought a new Marantz amp. It has since been lying around with me for a couple of years. The past month I got an itch to feel the smell of the solder again, so I opened it up to see the internals. I found a few burnt resistors and after doing some research, I replaced the resistors and turned on the amp again. I was amazed to find that's all that was wrong with the other channel, which was now working with no issues. I was also very impressed with the sound quality coming out of it. This made me think if I could improve it further.
The first step was to figure out why the resistors got burnt, but the answer was right in front of me in the Quad 306 service manual. Page 10, note (C) - "The centre-tapped DC Line is floating and some faults may cause this to shift putting up to 80v on one side possibly damaging R30/R31 (2K2ohm)." However, I also had R22 (560R) smoked out, not sure why.
So after replacing every single resistor with better-specked Vishay resistors and doing some upgrades, I am gobsmacked with how good this thing sounds. I am literally in awe and just in love with it. Here is the list of things I have done;
- Main power capacitors upgraded to 6800uf 63v, bypassed by 0.1uf Panasonic film capacitors.
- Replace the rectifier bridge with GBU604 (Strictly speaking, not required)
- Bypass rectifier bridge outputs with 0.1uf 250v Panasonic capacitors to reduce high-frequency noise and diode ringing (TI - Power supply design guidance)
- Put a Schottky diode across the main power supply capacitors to clamp down the ground ( LesW - Quad 306, Pink Fish Media post)
- Put a 22uf decoupling capacitor across zener diode D1, and D2 to filter the power supply to TLC271 ( LesW - Quad 306, Pink Fish Media post)
- Put a 100nf decoupling capacitor across zener diodes D3 and D4. (Dada upgrade guidance)
- Replace the existing 120k R6 resistor with 62k resistor (Dada upgrade guidance)
- Replace the existing 9R1 resistor R13 to 27R resistor to change input sensitivity to 1.0V (Dada upgrade guidance)
- Replace existing 47uf capacitor C7 with 100uf 63v capacitor (Dada upgrade guidance)
- Capacitors C1, C4, C5, and C6 were replaced like for like with polystyrene capacitors. (Dada upgrade guidance)
- C2 replaced with 330nF and C3 replaced with 1uf MKT capacitors (Dada upgrade guidance)
- Capacitor C8 47p Silver mica was replaced with a brand new Silver mica capacitor, like for like. (Dada upgrade recommends polystyrene here)
- Replaced all zener diodes and 1n4003 with the latest versions like for like.
- Replaced TLC271 with a newer version. TLC271CP (original) is rated at 0 to 70 degrees temperature and TLC271IP is rated at -40 to 85 degrees temperature. For the price of another 2 euros, I thought why not. Plus after 25-30 years of technological improvements, I am sure the latest ones are better in every aspect of performance, even when working as a DC servo.
I noticed, since putting Schottky diodes across the main capacitors the amp doesn't heat up much at all. It has most certainly brought the working temperatures down a lot. At low volumes, it doesn't even get warm and stays cold to the touch. I can have the amp running for hours and it barely gets warm. Whereas before it use to get quite a bit warm/hot to the touch. The difference in heat reduction is huge.
Few warnings for anyone who might be working on these little gems
- There are different revisions to the PCB and minor improvements along the way. Check the PCB revision and the corresponding schematics before you tweak any values.
- On the Revision 2 PCB the marked direction of the diode D2 is incorrect, so be careful. On Revision 3 PCB this was corrected.
- R12 is marked as 2K2 on schematics, but 330 ohms is installed. On Revision 3 schematics this was corrected to show 330 ohms.
- Use Quad 306 schematics as guidance only and double-check the value of resistors installed in the amp. Replace like for like.
- After doing these upgrades/changes the T3 emitter voltage is about 60mV (schematics look for 16mV)
- TLC271 output on schematics is looking for -111v but I am getting -20mV and - 5mV here. This is normal and fine, please read the explanation by @Rupopulles in the comments below.
All in all, I am still in shock by how good this amp sounds, I am hearing instruments in the songs, which I didn't even know existed before. Now I am wondering if my crappy speakers sound so good with it, how would it sound with decent quality speakers. Looks like it's about time to DIY a 2-way speaker system. This is it, folks, I thought I might share my results just in case it might help anyone else here. Happy DIYing and thanks to you all who contribute to the forums, it helps the likes of me (Civil Engineer by profession but with a DIY audio bug).
Attachments
Last edited:
Quad's 306 and related models do, even in original form, have an uncanny, musical sound quality and staging that is lacking in most conventional design amps. One reason is l think, that like all of Quad's SS amplifiers of the period, it lacks the ubiquitous LTP input stage. Whilst selectively cutting 2nd harmonic distortion does reduce amplifier THD significantly, it also has the effect of skewing the remaining distortion products to the odd and less pleasant harmonic side and this isn't a good thing, but we seem to accept that when this THD is low enough, it's also quite acceptable. Still, I'm quite happy with the "singleton" input stage and even some variations to the "long tailed pair" circuit that alter the harmonic mix, even though this results in higher distortion.
That LTP problem intrigues me as it remains a feature of just about every other linear audio amplifier design out there. I guess it shows that like an efficient workman, designers go for convention and the biggest hammer first, knowing it will have the most impressive effect on the overall numbers, whatever the cost.
Last edited:
I know someone who had modified his Quad 306 and listened through Martin Logan CLS speakers without any issues.
My revamped Quad 306 is sounding way better than my friend's Marantz 6005. I don't know the ins and outs of the design (my lack of electronics background) but as you have explained it could be to do with the lack of input stage that creates its own issues. I seriously doubt I would be able to find another amp that sounds this good without spending a fortune. I have just ordered a new pairs of speaker terminals for the amp as my last update to it.
Cheers Hariharpal, My dad owns a 306. Last year I replaced the PSU caps for fresh 6m8u Kemet caps and he said it already helped sound and reduced hum.
I notice I have like 100 of the same panasonic caps youre using! Should I use those for bypassing in my dads 306?
Grazie Mille!
I notice I have like 100 of the same panasonic caps youre using! Should I use those for bypassing in my dads 306?
Grazie Mille!
The PinkfishMedia thing about the diodes across the PSU caps is nonsense. They don't do what is claimed, and if they did they would be fighting with virtual-earth transistors T11/12 and probably blowing them up, and R30/31 as well.
If you have wrong currents and voltages you still have a fault.
If you have wrong currents and voltages you still have a fault.
Absolutely, you have nothing to lose, it's very common practice among the audio DIY community. I would advise reading the TII power supply design guide "AN-1849 An Audio Amplifier Power Supply Design", just for the sake of knowledge and understanding.
ejp: thanks, but it's not just the forums from that website, there are numerous other designs using the same trick. As far as the voltages go, we are talking in mV category, and after doing dada upgrades and changing all the resistors to 1% I doubt I will be able to match it. The main thing for me is that there is no DC on the speaker terminals and the Amp sings like an opera singer. So I am going to leave it at that, if anything was wrong with what I did it would have shown in the music production.
If you want equal voltages across two capacitors in series you need equal resistors shunting them, of a low enough value that some current can actually flow. 1M is typical. A diode has far too much impedance in the reverse direction to accomplish this feat: and in any case the virtual-earth in the 306 is designed to produce unequal voltages. That's why R28/29 have different values,and that's why the rails are labelled +40.6V and -37.9V. Ergo, if you haven't blown up the virtual-earth circuit, the diodes are not doing what is claimed by LesW. What they are really doing is providing protection against flyback, but the correct way to do that is to connect them between the rails and the output point, across the output transistors.
Hi thanks, for explanation. My understanding is that the whole idea of diodes across the capacitors is to protect against flyback and nothing else. Going with what was mentioned in the Quad 306 service manual about the floating power supply and how it could damage the resistors R30/R31 in case of voltage swings (which was the case with this amp), I decided to look for solutions. I will investigate this further as per your recommendation, thanks.
In regard to mV voltages at 2 points, I don't know what else could be wrong as all other voltages match near enough. Though my main power rails are about 2V less on either side +38 - 36, and it has been that way from the start. On the other note there are few resistor values that were different from the schematics, so I am wondering if that mV values could be trusted. And then again how come everything is working like a charm. I mean even at full volumes there is no distortion, hum or hiss?
In regard to mV voltages at 2 points, I don't know what else could be wrong as all other voltages match near enough. Though my main power rails are about 2V less on either side +38 - 36, and it has been that way from the start. On the other note there are few resistor values that were different from the schematics, so I am wondering if that mV values could be trusted. And then again how come everything is working like a charm. I mean even at full volumes there is no distortion, hum or hiss?
Ok. I had a look around with the voltages again today. All voltages are bang on to what is mentioned in the schematics, what's not said in the schematics is power rails are to be tested under load to match the said values.
The only voltages that don't match are -22mV and -5mV on either channel at TLC271 output (-111mV on Schematics) and T3 and C4 junction at 59mV (16mV on schematics).
Could this be because I have changed the resistor value R6 with 62k (as per dada upgrades), which is 120k on schematics? R13 changed to 27R to increase the input sensitivity to 1.0v. I am not sure as I have limited electronics knowledge. However, the proof is in the pudding, with no DC on the output terminals, no hum, and no distortion. The amp has been running for hours each day and I have no issues with it at all.
The only voltages that don't match are -22mV and -5mV on either channel at TLC271 output (-111mV on Schematics) and T3 and C4 junction at 59mV (16mV on schematics).
Could this be because I have changed the resistor value R6 with 62k (as per dada upgrades), which is 120k on schematics? R13 changed to 27R to increase the input sensitivity to 1.0v. I am not sure as I have limited electronics knowledge. However, the proof is in the pudding, with no DC on the output terminals, no hum, and no distortion. The amp has been running for hours each day and I have no issues with it at all.
That's not what LesW says at pinkfish. He claims it equalises the voltages and cures 'vague reproduction', whatever that means, and also provides a low impedance return path, which is BS. Your measurements prove him wrong, and self-deluding. I don't know why Dada recommended 62k. I never do it.
Hi Hariharpal,
Please share the schematic youre talking about in pdf or png. With quick acces ill be happy to have a look.
Dont worry or act on those who give negative input. Everybody deserves a chance to learn.
Much cheers and happy listening,
Ruben
Hi Ruben,
I have updated my original post with the schematic showing the dada upgrade values and voltage readings that are different. Please keep in mind that the amplifier is working perfectly fine and has no DC on the speaker terminals. I am just confused, as to why these voltages are not matching the service manual values. Every other voltage reading is exactly the same as per the schematics. Thanks.
Hari
I have updated my original post with the schematic showing the dada upgrade values and voltage readings that are different. Please keep in mind that the amplifier is working perfectly fine and has no DC on the speaker terminals. I am just confused, as to why these voltages are not matching the service manual values. Every other voltage reading is exactly the same as per the schematics. Thanks.
Hari
Oh the beauty of the 306 design. As a thrid year EEE student im still trying to understand how it performs feedfoward. In my eyes 306 current dumping is one of the most intriguing designs out there. Thanks for adding the notes too. Very clear.
Youre on the right track. The TLC271 operates as a DC servo to null out any DC at the output. As you mentioned it clearly works. Since youre using 62k, the opamp's output doesnt need to go as low to achieve the same adjustment. It also adjusts dc levels to counter component variations, hence the difference between the channels.
T3 is loaded by a current source, so changing to 27ohms should result in a higher voltage over R13. I assume you learned about ohms law.
My knowledge does tell me however that changing R13 from 9.1 to 27 ohms decreases input sensitivity and voltage gain. It was 0.375V now it is 1V. You have now reduced votlage gain but increased loopgain.
Btw who can tell me what the voltage gain is exactly. Its not quite R22/R13 + 1 is it?
Cheers!
Ruben
Youre on the right track. The TLC271 operates as a DC servo to null out any DC at the output. As you mentioned it clearly works. Since youre using 62k, the opamp's output doesnt need to go as low to achieve the same adjustment. It also adjusts dc levels to counter component variations, hence the difference between the channels.
T3 is loaded by a current source, so changing to 27ohms should result in a higher voltage over R13. I assume you learned about ohms law.
My knowledge does tell me however that changing R13 from 9.1 to 27 ohms decreases input sensitivity and voltage gain. It was 0.375V now it is 1V. You have now reduced votlage gain but increased loopgain.
Btw who can tell me what the voltage gain is exactly. Its not quite R22/R13 + 1 is it?
Cheers!
Ruben
Ruben, many thanks for the explanation. It all makes sense now. As, I am using this amp with my computer connected to usb DAC, I have to reduce the senstivity to 1.0v (used Dada upgrade guidance for correct resister values), it is now less aggressive with DAC volume.
I am sure someone more knowledgeable will be able to answer your voltage gain question. I am absolutely delighted with this amplifier.
Thanks again.
Hari.
I am sure someone more knowledgeable will be able to answer your voltage gain question. I am absolutely delighted with this amplifier.
Thanks again.
Hari.
Last edited:
I've just bought a 306 with broken LS sockets for a good price. I've owned one before a few years ago and this one sounds as good.
I've ordered the parts recommended by DADA and upgraded psu caps (10,000 50v).
However I am puzzled as to why the Schottky diodes would reduce dissipation as this is surely due to the class A standing current? Reducing this would affect the sound quality, you'd think or Quad would have done it.
I've ordered the parts recommended by DADA and upgraded psu caps (10,000 50v).
However I am puzzled as to why the Schottky diodes would reduce dissipation as this is surely due to the class A standing current? Reducing this would affect the sound quality, you'd think or Quad would have done it.