• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Commercial Gainclone kit- building instructions

Hi folks, been lurking for quite some time and ive finally got to the point where I can put this together, Ive been collecting bits and after reading the entire thread I still have some questions, i hope someone can help....
Ive got a large case been delivered monday from HiFi2000, Ive got 2x 300va 25v transformers and I would like to keep the chips running as cool as pos so I have 3 options, 10mm L shaped aluminium angle in 2 sizes: 63mm x 63mm x 250mm long (1 of) and 102mm x 77mm x 250mm long. (2 of) Third choice is something i was saving for my DCB1 build but I dont know if this amp really needs it so I could use them on the chips. 2x heat sink of 4mm back plate, 230mm L x 103mm H with fins 40mm. They will all fit inside the case, probably overkill but Its all bits I have spare. Maybe some pics would be better, I will try and upload a few. Thats my first hurdle.....
20151211_233846.jpg
[/URL][/IMG]

20151211_233823.jpg
[/URL][/IMG]

20151211_233745.jpg
[/URL][/IMG]

I should of said I have 4 amp boards, 2 where donated so I plan to use them to drive my sub amps, I would really like to know, do I really need to drive my sub amps from a second pair of 3875's? I have seen the question asked and the answer was a pair of 3875's will drive sub amps under certain conditions, IE the type of fets on the input, I figured Ive got 4 so it may be better? So I will potentially mount 2x 3875's per Sink. this would rule out the smaller angle.
I will save my other questions for later. :eek:
 
Last edited:
The solid aluminium angle has much mass and so will slow down changes of temperature in the chip back plate.
It has a low surface area relative to it's mass. This will determine how far above ambient it operates with the average power dissipated into it.
The aluminium angles are both very thick. This will very effectively take heat out to the extremities, i.e. the whole angle will be near the same temperature.

The finned heatsink is lowish mass and so will change temperature more quickly.
It is high surface area, that will dissipate the average heat more readily.
The back plate is fairly thin in comparison to the major dimension. This will result in the extremities operating at a much cooler temperature relative to the interface contact area. This thin back plate would suit lots of devices spread across the interface area, rather than one or two chipamps located near the middle.

Use the thick aluminium angle, with the finned heatsink, to reduce the aluminium temperature.
And ensure the heat can get out of the enclosure.
 
The solid aluminium angle has much mass and so will slow down changes of temperature in the chip back plate.
It has a low surface area relative to it's mass. This will determine how far above ambient it operates with the average power dissipated into it.
The aluminium angles are both very thick. This will very effectively take heat out to the extremities, i.e. the whole angle will be near the same temperature.

The finned heatsink is lowish mass and so will change temperature more quickly.
It is high surface area, that will dissipate the average heat more readily.
The back plate is fairly thin in comparison to the major dimension. This will result in the extremities operating at a much cooler temperature relative to the interface contact area. This thin back plate would suit lots of devices spread across the interface area, rather than one or two chipamps located near the middle.

Use the thick aluminium angle, with the finned heatsink, to reduce the aluminium temperature.
And ensure the heat can get out of the enclosure.


Many thanks for the info and for replying so quick Andrew, I now know more about heat sinks than I ever did. :D You are spot on with the finned, it had 10 chips on it, I never thought of using both the angle and the fins but they where made for each other.
URL]


This would definitely rule out using the DCB1 but unless someone tells me i NEED it Im happy. I dont really want to change it once Ive completed, I dont have the time or the inclination these days. So that leaves the question of wether i really need 4 3875's? Originally I planned to split the DCB1 output but Ive got options now, my volume control, a BtfSystem LDR will be in the same enclosure so I doubt the buffer will be required, the main source will be a usb Dac and the front channels of my AV system.
 
I recommend a Buffer where the Source does not have the current capability to drive the interconnect cable that follows.
This requires the Buffer at the Source Output.
The Buffer does not do any good at the Input of the Receiver.

Except when driving the ultra low input impedance of an inverting receiver.
Non inverting Power Amplifiers typically have input impedances of 10k to 1M
Inverting Power Amplifiers can have input impedances of 300r to 10k
 
Last edited:
I recommend a Buffer where the Source does not have the current capability to drive the interconnect cable that follows.
This requires the Buffer at the Source Output.
The Buffer does not do any good at the Input of the Receiver.

Except when driving the ultra low input impedance of an inverting receiver.
Non inverting Power Amplifiers typically have input impedances of 10k to 1M
Inverting Power Amplifiers can have input impedances of 300r to 10k

Cheers Andrew, that definitely rules it out, I will hang on to it, the 3875's might not be my cup of tea.
 
Check few pages back: http://www.diyaudio.com/forums/audi...it-building-instructions-101.html#post4459777

Another forum member uses similar case without any heatsinks or angles and it seem to work fine.

I would attach the chips to the front panel directly and use some angles to connect front and sides for improve heat transfer. In that way the chassis becomes one big heatsink and it should be enough to dissipate all the heat.

While what Andrew says earlier is all valid, it relates more to a typical SS amps, which Gainclones not really are. I've built numerous amps in some weird enclosures with very little metal parts and it all proved to work correctly.

For instance: 6 channels in an enclosure much smaller than yours.
 

Attachments

  • p5.jpg
    p5.jpg
    196.5 KB · Views: 303
Cheers Peter, I was just going on what I've read with the higher heat from higher voltages. Bolting it to the front plate sounds like a good idea but I won't be able to sleep knowing I've got longer signal wires, the case is quite big, I wanted it to be a chassis for all occasions. Still, i dont want any unnecessary bits of iron i dont need. What do you think about running multiple 3875 boards? I would like to keep it simple but I don't want to connect my powered subs to a single board if it could compromise the performance. The spare boards came from a friend who has decided not to build the amp.
 
The heat is mostly generated by speaker loads and how loud you play, higher supply voltage does not really matter if you don't push the amp.

I wouldn't like long signal wires neither, so why not attaching the angles to side panels, You may add additional small angles to the front.

If you plan on subwoofer, using separate channels for that purpose is definitely a good idea, it also doesn't need to be a premium version.

Here's an example of another design, with two main channels and two sub channels.

I also used a transformer with two separate sets of windings: 2 x 20V for main and 2 x 25V for subs. This allowed me to separate PS components: 100uF BG N for main channels and much more capacitance for sub channels. Not mixing those two bank of caps was very critical to achieve specific tuning of the sound.
 

Attachments

  • a2.jpg
    a2.jpg
    65.8 KB · Views: 199
  • a1.jpg
    a1.jpg
    49.3 KB · Views: 198
  • 11.jpg
    11.jpg
    212.2 KB · Views: 309
  • 9.jpg
    9.jpg
    238.8 KB · Views: 293
  • 6.jpg
    6.jpg
    126.3 KB · Views: 305
  • 4.jpg
    4.jpg
    163.8 KB · Views: 309
The heat is mostly generated by speaker loads and how loud you play, higher supply voltage does not really matter if you don't push the amp.

I wouldn't like long signal wires neither, so why not attaching the angles to side panels, You may add additional small angles to the front.

If you plan on subwoofer, using separate channels for that purpose is definitely a good idea, it also doesn't need to be a premium version.

Here's an example of another design, with two main channels and two sub channels.

I also used a transformer with two separate sets of windings: 2 x 20V for main and 2 x 25V for subs. This allowed me to separate PS components: 100uF BG N for main channels and much more capacitance for sub channels. Not mixing those two bank of caps was very critical to achieve specific tuning of the sound.

Hifi porn!!!!!!
So from what your saying I should really do 2 stereo sets rather than dual mono? This would allow for a separate transformer for the subs. Ive already got 4 premium boards so thats what I will use, do I need more capacitance for the subs when they have their own sub amps? I have a pair of Hawthorn Audio Silver Iris Duets with dedicated amps for the LF. Oh and I play movies loud so the chips will be driven hard!
I probably wont get a chance to check back tonight but your input is very much appreciated Peter.
 
The main, full range channels should have very low capacitance with quality caps. BG N 100uF are perfect for that, but they are expensive and hard to find. Some people complain that the amps might not have enough "juice" for the subwoofers, that's why in that case I suggest more capacitance, up to 10,000uF, but only if the main and sub channels are separated, each having dedicated power supply.

Here's one other amp, I've built for my personal use powering Altecs. It is completely balanced with LM3875 for full range and LM4780 for woofers. I didn't add any extra capacitance in low freq channel as the speakers are very efficient.

BTW, the heatsink gets barely warm and those are monoblocks.
 

Attachments

  • _DSC2389.jpg
    _DSC2389.jpg
    124.4 KB · Views: 205
  • _DSC2392.jpg
    _DSC2392.jpg
    93.4 KB · Views: 186
  • _DSC2394.jpg
    _DSC2394.jpg
    83.6 KB · Views: 156
  • _DSC2395.jpg
    _DSC2395.jpg
    90.8 KB · Views: 156
  • _DSC2400.jpg
    _DSC2400.jpg
    191.6 KB · Views: 164
Last edited:
The main, full range channels should have very low capacitance with quality caps. BG N 100uF are perfect for that, but they are expensive and hard to find. Some people complain that the amps might not have enough "juice" for the subwoofers, that's why in that case I suggest more capacitance, up to 10,000uF, but only if the main and sub channels are separated, each having dedicated power supply.

Here's one other amp, I've built for my personal use powering Altecs. It is completely balanced with LM3875 for full range and LM4780 for woofers. I didn't add any extra capacitance in low freq channel as the speakers are very efficient.

BTW, the heatsink gets barely warm and those are monoblocks.
I like the vertical arrangement. The slim enclosures can be stacked side by side for a neat installation. Or stood behind a speaker and not take up much space.
 
The main, full range channels should have very low capacitance with quality caps. BG N 100uF are perfect for that, but they are expensive and hard to find. Some people complain that the amps might not have enough "juice" for the subwoofers, that's why in that case I suggest more capacitance, up to 10,000uF, but only if the main and sub channels are separated, each having dedicated power supply.

Here's one other amp, I've built for my personal use powering Altecs. It is completely balanced with LM3875 for full range and LM4780 for woofers. I didn't add any extra capacitance in low freq channel as the speakers are very efficient.

BTW, the heatsink gets barely warm and those are monoblocks.

These are all great designs Peter, makes me feel bad for just ordering a case, where do you find the time!

I did a search for subwoofer on this thread and it only returned 12 results, only about 2 of these where connection related and from your replies it looks like I have many options, in one post you say connecting a sub to the speaker out is fine, that's good to know, in another you said splitting the signal at the volume pot is fine, also good to know, I will contact Karl to see if his LDR control would be happy doing this. Now we have this other method which is probably the 'best', If the volume pot is compromised by a split I think I will order 1 more 25v transformer and run a pair of 3875's in stereo for subs, (my chassis will accommodate) this way I will keep the channel separation of dual mono on my full range drivers and also keep the subs power separate. I guess it will all be down to how well the LDR can drive long-ish line level to my sub amp locations. I may now need to find some more caps!
 
Well I've got some more info Peter, splitting the LDR between either 4 3875 boards or the sub amps and the 3875s is going to create an impedance mis match, or at least an un ideal load for the pot. So my next question is, can I change the input impedance of the amps to say 50k? I realise I could be asking a really stupid question but I'm learning as I go, obviously there is an ideal operating point for the amp and it's 22k for a reason, I am now looking for the best compromise, not a word I enjoy using but I have to accept my speakers are not your average set up, the subs are not really subs, they are an active LF portion of a 3 way unit, bass augmentation rather than subwoofers, so even though there 15" cones they are not designed to do really deep bass. I need to make some purchasing decisions quick if I want to get everything for the holiday.