Gainclone makeover and measurements.

Systems
Systems Posts: 14,873
edited December 2008 in DIY, Mods & Tweaks
This project started as a dead Sherbourne 2/100 amplifier.Instead of repairing it I thought it would be an ideal platform for a pair of LM3875 chip amps. I stripped it clean keeping only the transformer,chassis,AC inlet and power switch.The pics show some before and after shots.

As it turns out ,the transformers voltage and current capabilities are just about perfect(the rectified DC voltage is -+36)and the complete top and sides are heatsink which is more than sufficient to keep the chips operating well within their thermal limits.In fact it has enough surface area that even 4 of the LM3875's should run cool enough.I would have used 4 but the transformer cannot supply enough current to run more than 2.

The makeover was very simple the hardest part was making the new power and output harnesses.
I mounted the new bridge rectifier and filter caps on to bottom panel with the transformer.I salvaged the cap board from an old component but installed new 6800 uf caps.
I would have prefered to use a pair of 10,000uf but space didn't allow it.Some build their Gainclones with as little as 1500uf but for a number of reasons I think the more filter capacitance the better.

The transformer was a bit noisy as a result of it vibrating the thin guage panel it was mounted to.In an atempt to reduce the vibration I installed some sticky backed foam underneath it which was quite effective.

I decided to use input coupling caps to block DC from the source and used some 10k pots as I needed independant level adjustment for each channel.


I had access to an AudioPrecision analyzer so I did some testing,the results of which I will post soon.
Testing
Testing
Testing
Post edited by Unknown User on

Comments

  • ben62670
    ben62670 Posts: 15,969
    edited August 2008
    What voltage rails are you running? Love the case. Are those Brian's boards?
    Please. Please contact me a ben62670 @ yahoo.com. Make sure to include who you are, and you are from Polk so I don't delete your email. Also I am now physically unable to work on any projects. If you need help let these guys know. There are many people who will help if you let them know where you are.
    Thanks
    Ben
  • Systems
    Systems Posts: 14,873
    edited September 2008
    These test results were acquired from an Audio Precision portable One analyzer. The LM3875's were operated from +- 36 volt DC rails and configured for 27db of voltage gain.




    THD
    30 Watts into 8ohms 2k left .0031 right .0023

    20k left .036 right .030


    IMD at 30 watts left .0024 right .0025


    S/N
    unweighted 10hz-30k left 99.8 right 100.2


    input sensitivity left 783 mv right 783mv
    for 40 watts out


    Power output
    1% clipping at 2k 8 ohms left 44.2 w right 44.5w

    4 ohms left 49w right 49w


    DC offset left 53 mv right 73mv



    Other than offset both chips are very closely matched.The offset is rather high but this is typical for these devices.

    Both THD and IMD distortion is admirably low up to about 30 watts or so but climbs dramatically as it approaches 40 watts.With the 36 volt rails I would rate this as a 44 watt amp and best suited for 8 ohm loads.
    Testing
    Testing
    Testing
  • ben62670
    ben62670 Posts: 15,969
    edited September 2008
    Why did you select the 3875's over the 3886's?
    Please. Please contact me a ben62670 @ yahoo.com. Make sure to include who you are, and you are from Polk so I don't delete your email. Also I am now physically unable to work on any projects. If you need help let these guys know. There are many people who will help if you let them know where you are.
    Thanks
    Ben
  • Systems
    Systems Posts: 14,873
    edited September 2008
    ben62670 wrote: »
    Why did you select the 3875's over the 3886's?
    I bought a bunch of blank boards a couple of years ago from Audio sector,they did'nt stock 3886 boards. My understanding is that they are very close in performance.
    Testing
    Testing
    Testing
  • ben62670
    ben62670 Posts: 15,969
    edited September 2008
    The 3886's are 50/68 watts 8/4 ohms. Have you considered running another pair of those amps parallel with those? There wouldn't be much of a power gain, but it would handle the heat much better, and you could run 4ohms hard. With that voltage most would say you are pushing it a little too much. I have mine running the same as yours for now. All you need to tie them together is a couple .2r5w's on the outs;)
    Ben
    Please. Please contact me a ben62670 @ yahoo.com. Make sure to include who you are, and you are from Polk so I don't delete your email. Also I am now physically unable to work on any projects. If you need help let these guys know. There are many people who will help if you let them know where you are.
    Thanks
    Ben
  • Systems
    Systems Posts: 14,873
    edited September 2008
    ben62670 wrote: »
    The 3886's are 50/68 watts 8/4 ohms.
    Slightly more power but that would be dependant on rail voltages used for each.I believe the SQ to be close between them
    Have you considered running another pair of those amps parallel with those?There wouldn't be much of a power gain, but it would handle the heat much better, and you could run 4ohms hard.
    If I was running a 4 ohm load that was full range (not just a tweeter)then I would certainly consider paralleling them if my transformer was up to the task.Unfortunately it only has enough current for 2.
    With that voltage most would say you are pushing it a little too much.
    This voltage would only be a problem if the heatsinking was inadequate.
    Testing
    Testing
    Testing
  • ben62670
    ben62670 Posts: 15,969
    edited September 2008
    I am rung a single chip just for now at those voltages with no issues. We are right on the line according to the DIY crowd. What they are saying is that with a chip as powerful as these the chip itself can only dissipate so much heat to the heat sink. I don't know about that though. I drove mine very hard. A sine wave at 1khz 100watts for 15 seconds, and it did not shut down. I know the distortion figures must not have been pretty, but I did want to see what would happen. The amp chips are only $3.50 each :)
    Ben
    Please. Please contact me a ben62670 @ yahoo.com. Make sure to include who you are, and you are from Polk so I don't delete your email. Also I am now physically unable to work on any projects. If you need help let these guys know. There are many people who will help if you let them know where you are.
    Thanks
    Ben
  • Systems
    Systems Posts: 14,873
    edited September 2008
    ben62670 wrote: »
    We are right on the line according to the DIY crowd.
    Yeah wouldn't want to go much higher even though the data sheets say it's do able.
    I drove mine very hard. A sine wave at 1khz 100watts for 15 seconds, and it did not shut down.
    Wow really,the protection circuits in these seems to be rather agressive so its surprising it didn't shut down.
    The amp chips are only $3.50 each :)
    My chips must be better or else I got hosed,as they were $7 a pop.:D
    Testing
    Testing
    Testing
  • dudeinaroom
    dudeinaroom Posts: 3,609
    edited December 2008
    I wanna build one, I have a couple of schematics bookmarked for when I get some dough. I was thinking of going with 4 chips per channel bridged, in parallel so I can get higher power, and still be able to run a 4 ohm load. who knows, I might just end up with a cool smoke generator.
  • ben62670
    ben62670 Posts: 15,969
    edited December 2008
    Let me know if you are going to go for it. I have a line on a high amp transformer for youi.
    Ben
    Please. Please contact me a ben62670 @ yahoo.com. Make sure to include who you are, and you are from Polk so I don't delete your email. Also I am now physically unable to work on any projects. If you need help let these guys know. There are many people who will help if you let them know where you are.
    Thanks
    Ben
  • dudeinaroom
    dudeinaroom Posts: 3,609
    edited December 2008
    All depends on Cash flow. Gotta wait and see how every thing pans out in Feb, and If I get an HDTV, or just spring for a converter box
  • ben62670
    ben62670 Posts: 15,969
    edited December 2008
    Living the high life there:) I have a 19" TV with rabbit ears. I get one station, and I am fine with it. When it goes digital I will have to spring for the box:mad:
    Please. Please contact me a ben62670 @ yahoo.com. Make sure to include who you are, and you are from Polk so I don't delete your email. Also I am now physically unable to work on any projects. If you need help let these guys know. There are many people who will help if you let them know where you are.
    Thanks
    Ben
  • fredv
    fredv Posts: 923
    edited December 2008
    I wanna build one, I have a couple of schematics bookmarked for when I get some dough. I was thinking of going with 4 chips per channel bridged, in parallel so I can get higher power, and still be able to run a 4 ohm load. who knows, I might just end up with a cool smoke generator.
    I am waiting for my boards from this group buy ........

    -fredv-
  • dudeinaroom
    dudeinaroom Posts: 3,609
    edited December 2008
    ben62670 wrote: »
    Living the high life there:) I have a 19" TV with rabbit ears. I get one station, and I am fine with it. When it goes digital I will have to spring for the box:mad:

    goto dtvanswers.com and apply for one of the 40 buck off coupons. Some places on the net have them for $40 and free shipping. I applied for mine yesterday.