DC offset issues with GC

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
i built a GC per Peter Daniels schematic with 1K input resistor. the schematic is on the first page: http://www.diyaudio.com/forums/audi...cial-gainclone-kit-building-instructions.html

with the inputs shorted, i get DC offset of (L)6.9mV and (R)0.3mV. with both inputs floating, i get around (L)22mV and (R)26.

my preamp has a DC offset of (L)0.5mV and (R)0mV.

when i connect the pre to the GC (no source connected), DC offset jumps to more than 500mV.

i get very little DC offset with a Benchmark DAC1 or an iPod connected to the GC.

any ideas?
 
embarrassing as this is, i should also add that last week i fried a brand new pair off Fostex FF85K (10W max) when powered with the GC and iPod in a 24'x24' garage, after a "few" beers with a friend, and Weezer cranked way too loud. all my DC measurements were taken today.
 
First of all you should turn to Peter Daniel since he have sold you this product, still something isn't right. Is the preamp also homebrewed? What is the DC voltage from the preamp when it's connected to the Gainclone?

I think Peter should add some proper start-up procedure in his documention. How about inserting a small fuse? :idea: 100 mA when you are testing and/or a 2200 uF cap.

Power ground and signal ground together?
 
Last edited:
this is not a PD product. i only used his schematic and don't wish my DIY efforts to reflect poorly of him. here's some pics from another thread detailing some instability issues i had, which were sorted: http://www.diyaudio.com/forums/chip-amps/147135-gc-noise-heat-issues.html

DC offset at the preamp is less than 1mV in both (L)(R) when connected to the GC.

both signal and power are tied to ground.

the preamp is detailed in the above linked thread. it's a actually a headphone amp serving both preamp and HP amp duties, and has an on-board DAC regulated at 5V by an LM317. all of my DC offset measurements were with no transport connected to the DAC. the headphone amp works fine in isolation.
 
R1= 1K
R2= 22K
R3= 1.8K (gain=13 dB, where 1+Rfb/R3=gain)
Rfb= 22K
Cs at V+ and V-= 1500uF

details of pictures:
-the twisted/stripped wire bridging the 2 terminal strips at the 2 outside end tabs chip is ground.
-white wires (from PS) carry PG+ and PG- and connect to bridged wire/ground at the ends terminal strip, at opposite ends.
-red wires connect the terminal strips ground to star ground near the IEC/chassis
-black wires (from PS) carry V+ and V- and connects with each Cs at the 2 center tabs of the terminal strips and to pins 1 and 4 of the chip. the other Cs lead connects to ground at the terminal strip end.
-R2, R3, IG and OG (white wire directly to (-) speaker binding post) are connected directly to the bridged wire/ground, in the middle, while the PG connections are at the ends of the terminal strips.
-R1 is directly in series with red input wire and covered with heatshrink.

i tried removing R1 and placing a 4.7uF non-polar cap in-series with input and measured a higher DC offset at 1.4V. right now, R1 is back in-series, with the coupling cap removed.

thanks!

042.jpg
[/IMG]

043.jpg
[/IMG]
 
I suspect your pre-amp is home made too and does not include a DC blocking capacitor on its output. Then when you connect your amp, the DC offsets are all multiplied by the amp which also has no DC blocking or attention to DC conditions whatsoever, and produces an unacceptable output offset.

i'm having a difficult understanding why this would be when there is nil offset at the preamp. The Cavalli-Kan Kumisa III Stereo Headphone Amplifier

AndrewT,
i have read it, more than once. i'm no EE and have been DIYing for a little over a year. i guess i should interpret your post as "implement the chip as it was designed to"?
 
Just re-read your post and I missed the bit where you say you replaced R1 with a cap and it was worse. Also looking at your pre-amp it has an offset servo so should be ok.

I think your problem is the one that many builders of this particular gainclone seem to be having and that is the lack of DC gain roll off capacitor between R3 and ground. Try 100uF.
 
Just re-read your post and I missed the bit where you say you replaced R1 with a cap and it was worse. Also looking at your pre-amp it has an offset servo so should be ok.

I think your problem is the one that many builders of this particular gainclone seem to be having and that is the lack of DC gain roll off capacitor between R3 and ground. Try 100uF.
thank you - that was helpful. i'll give it a shot.
 
no bi-polar caps in the parts bin, but i found two 100uF polar caps i soldered together in series (+ to + = 50uF) and soldered these in series from R3 to ground in one channel only. DC offset dropped to 35 mV :). i'm placing an order for 47uF and 100uF bi-polars. i'll go with the 100uF if i can get them to fit with a reasonable amount of security with P2P.

thanks :spin:!!!
 
polar electrolytics will do the job as DC block in the NFB route.

Add a 1n4*** diode across the cap pins to prevent excessive reverse DC voltage.

If the amplifier is set up correctly there should never be any significant AC voltage across the cap.

The DC voltage will be the input offset voltage if the front end is reasonably well balanced. Check that the cap is oriented correctly to take this permanent DC voltage. Except when the output fails to one rail or the other, then the NFB resistors and the cap/diode share the supply rail voltage.
 
polar electrolytics will do the job as DC block in the NFB route.

Add a 1n4*** diode across the cap pins to prevent excessive reverse DC voltage.

If the amplifier is set up correctly there should never be any significant AC voltage across the cap.

The DC voltage will be the input offset voltage if the front end is reasonably well balanced. Check that the cap is oriented correctly to take this permanent DC voltage. Except when the output fails to one rail or the other, then the NFB resistors and the cap/diode share the supply rail voltage.

sorry, i'm a little confused. are you suggesting i can use a single polarized cap with the (-) to ground as long as i have a diode across the cap with it's cathode tied to ground?
 
the NFB DC blocking cap can be any type of low voltage or high voltage capacitor.

If the amplifier components have been selected correctly that DC blocking cap should never see any significant AC voltage across it. That means a polarised capacitor can be used.

If the amplifier is working properly then there should be a small DC voltage across it. This voltage is approximately equal to the input offset voltage, maybe in the range 5mV to 500mV. This voltage can be of either polarity. The input stage devices determine whether it is +ve or -ve. Check your voltage and orient the cap appropriately.

It is possible for the amplifier in exceptional/fault situations for the DC voltage across the cap to exceed 1V. Using a diode to prevent this reverse voltage removes the possibility of damaging the cap due to reverse voltage.
If you choose to use a capacitor of lower voltage rating than one of the supply rails then it's wise to also add a second diode in inverse parallel to the first. These two diodes prevent voltages across the capacitor exceeding 1V, even in full output offset (failed/shorted output stage).

Use a polarised capacitor sized to pass a wider bandwidth than the input DC blocking capacitor and select 10V to 16V working rating. These are cheap and tiny.
You can experiment by adding an extra capacitor across the electrolytic. This may /maynot improve the sound quality.
 
Last edited:
the NFB DC blocking cap can be any type of low voltage or high voltage capacitor.

If the amplifier components have been selected correctly that DC blocking cap should never see any significant AC voltage across it. That means a polarised capacitor can be used.

If the amplifier is working properly then there should be a small DC voltage across it. This voltage is approximately equal to the input offset voltage, maybe in the range 5mV to 500mV. This voltage can be of either polarity. The input stage devices determine whether it is +ve or -ve. Check your voltage and orient the cap appropriately.

It is possible for the amplifier in exceptional/fault situations for the DC voltage across the cap to exceed 1V. Using a diode to prevent this reverse voltage removes the possibility of damaging the cap due to reverse voltage.
If you choose to use a capacitor of lower voltage rating than one of the supply rails then it's wise to also add a second diode in inverse parallel to the first. These two diodes prevent voltages across the capacitor exceeding 1V, even in full output offset (failed/shorted output stage).

Use a polarised capacitor sized to pass a wider bandwidth than the input DC blocking capacitor and select 10V to 16V working rating. These are cheap and tiny.
You can experiment by adding an extra capacitor across the electrolytic. This may /maynot improve the sound quality.

thanks, i get it now :).
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.