Lindsay,
on the boards that I have, the third LED (front end ccs) is tied to FE ground (eyelet E2) thru a 22K1 resistor (22.1k ohms).
E2 is NOT the same ground as E7 (which feeds the 100uf BG rail RC filter caps only).
you may need to check LEDs with room lights off-- FE CCS LED are slightly dimmer than the other 2 leds on the boards i have.
-Bruce
on the boards that I have, the third LED (front end ccs) is tied to FE ground (eyelet E2) thru a 22K1 resistor (22.1k ohms).
E2 is NOT the same ground as E7 (which feeds the 100uf BG rail RC filter caps only).
you may need to check LEDs with room lights off-- FE CCS LED are slightly dimmer than the other 2 leds on the boards i have.
-Bruce
Bruce,
I can see the right channel LED under most conditions, so I am sure the left channel LED is either off or at considerably lower voltage.
I do not have a connection to either channel where the board shows chassis ground (near the input) and I have replaced the RCA connectors with ones that are isolated. Is it important to have a chassis ground to the board?
There is an eyelet connected to the chassis that is connected to the star ground. It was connected to the old RCA grounds, but per the MC instructions I removed the connection from the eyelet to the RCA ground lugs. During my testing the RCA input for the left channel is not connected to the board.
I received the matched MOSFETS from MC today so if the MC boards are still working I may get my amp back working this weekend. I'll post my results.
Lindsay
I can see the right channel LED under most conditions, so I am sure the left channel LED is either off or at considerably lower voltage.
I do not have a connection to either channel where the board shows chassis ground (near the input) and I have replaced the RCA connectors with ones that are isolated. Is it important to have a chassis ground to the board?
There is an eyelet connected to the chassis that is connected to the star ground. It was connected to the old RCA grounds, but per the MC instructions I removed the connection from the eyelet to the RCA ground lugs. During my testing the RCA input for the left channel is not connected to the board.
I received the matched MOSFETS from MC today so if the MC boards are still working I may get my amp back working this weekend. I'll post my results.
Lindsay
Lindsay,
from what i recall about the wiring on DH220, i set it up to use isolated grounds per John H's instruction (there are three wires connecting at/near RCA's to eyelets E1, E2, E3(not sure about numbering of this last eyelet, but its near the input section of PA3), and definitely a DC path from E2 back to the star ground at the junction of the large main filter caps (either its directly connected, or tied thru the 10ohm resistor on the board between E2 and e3)
this config helps reduce ground interaction between the two channels, if i recall my converstions with John H.
-Bruce
from what i recall about the wiring on DH220, i set it up to use isolated grounds per John H's instruction (there are three wires connecting at/near RCA's to eyelets E1, E2, E3(not sure about numbering of this last eyelet, but its near the input section of PA3), and definitely a DC path from E2 back to the star ground at the junction of the large main filter caps (either its directly connected, or tied thru the 10ohm resistor on the board between E2 and e3)
this config helps reduce ground interaction between the two channels, if i recall my converstions with John H.
-Bruce
Bruce et al,
I tested the original Hafler board with two different pairs of N and P MOSFETS. I Had the same result from each pair, that is a -2.5 VDC offset when adjusted to the minimum. I then tested the Hafler board with 47k resistors between the gate and source connections on the board with no MOSFETS connected at all I got a -3.5 VDC offset when adjusted to the minimum. I think this means the offset is a problem with the Hafler board. I also think the MOSFETS are OK since individually they seem to work and the offset appears to be caused by the board not the MOSFETS.
I did the same test on the MC boards with a 47k resistor from E6 to E8 and from E9 to E8. Both boards gave me -96.8 VDC at E8 (output) when the rails were +-98.0 VDC. From this I concluded that neither MC board is working. I checked closely for missing parts and the second board MC sent me is missing R1 and C2. Everything else looks the same between these boards and the board that is working.
I figured the only other thing to do was to check the resistors connected to the gates and they are right on. The capacitors I added across the MOSFETS are not shorted and seem to be OK.
I then put the new MOSFETS in with the Hafler board and again got -2.5 VDC offset when adjusted to minimum. I think the MOSFETS are fine and the MC boards are bad.
Opinions are welcome as are suggestions for additional tests. Thanks for any help.
Lindsay
I tested the original Hafler board with two different pairs of N and P MOSFETS. I Had the same result from each pair, that is a -2.5 VDC offset when adjusted to the minimum. I then tested the Hafler board with 47k resistors between the gate and source connections on the board with no MOSFETS connected at all I got a -3.5 VDC offset when adjusted to the minimum. I think this means the offset is a problem with the Hafler board. I also think the MOSFETS are OK since individually they seem to work and the offset appears to be caused by the board not the MOSFETS.
I did the same test on the MC boards with a 47k resistor from E6 to E8 and from E9 to E8. Both boards gave me -96.8 VDC at E8 (output) when the rails were +-98.0 VDC. From this I concluded that neither MC board is working. I checked closely for missing parts and the second board MC sent me is missing R1 and C2. Everything else looks the same between these boards and the board that is working.
I figured the only other thing to do was to check the resistors connected to the gates and they are right on. The capacitors I added across the MOSFETS are not shorted and seem to be OK.
I then put the new MOSFETS in with the Hafler board and again got -2.5 VDC offset when adjusted to minimum. I think the MOSFETS are fine and the MC boards are bad.
Opinions are welcome as are suggestions for additional tests. Thanks for any help.
Lindsay
fab,
I upgraded the power supply as part of this project. MC sent me new diodes and caps which I installed. The amp ran with the new parts for several weeks before the modification replacing the Hafler drive boards. As a result I think the power supply is OK. The only thing that gives me pause is that the voltage is 5 to 7 volts above the 90 volts that the amp was designed for.
That said, the right channel MC board appears to be working correctly, so I think the power supply is fine.
Thanks for the suggestion though. This is one of those problems that doesn't make sense. I would fine it a real quirk of fate that MC sent me two out of three boards that were defective. Although the fact that parts were missing from the latest board makes that look more possible. But it also appears that the MOSFETS are fine. There really isn't anything left to check.
Lindsay
I upgraded the power supply as part of this project. MC sent me new diodes and caps which I installed. The amp ran with the new parts for several weeks before the modification replacing the Hafler drive boards. As a result I think the power supply is OK. The only thing that gives me pause is that the voltage is 5 to 7 volts above the 90 volts that the amp was designed for.
That said, the right channel MC board appears to be working correctly, so I think the power supply is fine.
Thanks for the suggestion though. This is one of those problems that doesn't make sense. I would fine it a real quirk of fate that MC sent me two out of three boards that were defective. Although the fact that parts were missing from the latest board makes that look more possible. But it also appears that the MOSFETS are fine. There really isn't anything left to check.
Lindsay
I've had a couple of MC boards and noted a blank spot or two on them. Hillig includes places for components that in final trials may not be needed. He is a minimalist which leads to a cleaner sound. I doubt he would send out boards with important components missing and those blank spots represent the results of final testing.
If you measure voltages down the line on both rails, both sides, and find everything in spec then some board component/s must be the problem. I am astonished that you would get 97 VDC to the gates of the MOSFETs of one channel. Surely, this over voltage must also be appearing at some place on the PCB, probably on the emitters of the final driver transistors..
Do you have someone else around that you could have present as you talk through the connections? Sometimes we miss the obvious.
If you measure voltages down the line on both rails, both sides, and find everything in spec then some board component/s must be the problem. I am astonished that you would get 97 VDC to the gates of the MOSFETs of one channel. Surely, this over voltage must also be appearing at some place on the PCB, probably on the emitters of the final driver transistors..
Do you have someone else around that you could have present as you talk through the connections? Sometimes we miss the obvious.
Dick,
Thanks for sticking with me on this. The parts that were missing were only missing on the last board Hillig sent. It was missing R1 and C2 which I think are on the input, so I doubt these parts are the problem. The interesting thing is the voltage to the gates and outputs is approximately 1.2 VDC below (or above if you consider they are negative voltages) the rail voltage, which my brother says is the drop you would get across two semiconductor junctions. I have not made any measurements on the board as I really have no idea what to measure.
I know what you mean about missing the obvious. And it could be the problem here. When the left channel did not work the first thing I looked for was a wiring mistake. I did not have someone else help in checking the wiring, but I compared connection by connection with the board that was working. I am pretty anal about this stuff and I try and check the connections multiple times and in different orders. I don't believe I had a wiring mistake.
At this point I am pretty sure both MC boards have failed in the same way. I do not know what caused this, but I don't think it was the MOSFETS. Maybe the 97 to 98 VDC on the rails is right at the limit of what the MC boards can take. A weak component that would work at 90 VDC fails at 98 VDC. If I do any additional testing. I will thoroughly test with a light bulb in series with the power cord until I know everything is working. Or I may break down and get a Variac.
I've packed up the MC boards and will call Hillig in the morning and return them. He can then determine what failed and possibly then we will know why it failed. I'll post more as this project unfolds.
Lindsay
Thanks for sticking with me on this. The parts that were missing were only missing on the last board Hillig sent. It was missing R1 and C2 which I think are on the input, so I doubt these parts are the problem. The interesting thing is the voltage to the gates and outputs is approximately 1.2 VDC below (or above if you consider they are negative voltages) the rail voltage, which my brother says is the drop you would get across two semiconductor junctions. I have not made any measurements on the board as I really have no idea what to measure.
I know what you mean about missing the obvious. And it could be the problem here. When the left channel did not work the first thing I looked for was a wiring mistake. I did not have someone else help in checking the wiring, but I compared connection by connection with the board that was working. I am pretty anal about this stuff and I try and check the connections multiple times and in different orders. I don't believe I had a wiring mistake.
At this point I am pretty sure both MC boards have failed in the same way. I do not know what caused this, but I don't think it was the MOSFETS. Maybe the 97 to 98 VDC on the rails is right at the limit of what the MC boards can take. A weak component that would work at 90 VDC fails at 98 VDC. If I do any additional testing. I will thoroughly test with a light bulb in series with the power cord until I know everything is working. Or I may break down and get a Variac.
I've packed up the MC boards and will call Hillig in the morning and return them. He can then determine what failed and possibly then we will know why it failed. I'll post more as this project unfolds.
Lindsay
LHM,
when you get new boards back from MC, make sure that either E2 or E3 is connected to the star ground (junction of main filter caps), along with E8--I did some reviewing of the PCB layout in the MC manual, and the FE CCS needs E2/E3 tied to ground to work properly (this LED will be off or very dim).
glad to hear that the mosfets are OK. also, if the power supply is using the MC toriod upgrade, its likely that the rail voltages are higher than the 90V spec and normal for that toroid (on my DH-220, spec is about 60V, but the MC toriod for the DH-220 yields about 64V). Also, the schottky diode upgrade has slightly lower drop (about 1 volt) compared to conventional silicon junction diodes. I wouldn't be concerned about the supply rails.
Also consider the BG 100uF FK rail caps (~$80) for the PA3 boards as an additional upgrade.
Please keep us posted (wonderful sound awaits....)
-Bruce
when you get new boards back from MC, make sure that either E2 or E3 is connected to the star ground (junction of main filter caps), along with E8--I did some reviewing of the PCB layout in the MC manual, and the FE CCS needs E2/E3 tied to ground to work properly (this LED will be off or very dim).
glad to hear that the mosfets are OK. also, if the power supply is using the MC toriod upgrade, its likely that the rail voltages are higher than the 90V spec and normal for that toroid (on my DH-220, spec is about 60V, but the MC toriod for the DH-220 yields about 64V). Also, the schottky diode upgrade has slightly lower drop (about 1 volt) compared to conventional silicon junction diodes. I wouldn't be concerned about the supply rails.
Also consider the BG 100uF FK rail caps (~$80) for the PA3 boards as an additional upgrade.
Please keep us posted (wonderful sound awaits....)
-Bruce
LHM,
after re-reading my last post, I want to clarify a few things--
E2/E3 to ground is needed to bias FE CCS and its LED (if NOT properly grounded, then LED will be off or dim).
Rail voltage of ~97V should not be a concern (the MC electro FE rail caps are rated for 100V, and the MC film caps are also 100V rated or better).
Also, from my recollection, on stock DH-500 ground wiring is slightly different than for DH-220 (with L and R hooked up slightly differently).
I'm a bit concerned about one of your previous posts, where John H's instructions had you removing a ground wire (and not necessarily provding another grounding path). Double check with John about ground wiring details--I suspect something is amiss here.
-Bruce
after re-reading my last post, I want to clarify a few things--
E2/E3 to ground is needed to bias FE CCS and its LED (if NOT properly grounded, then LED will be off or dim).
Rail voltage of ~97V should not be a concern (the MC electro FE rail caps are rated for 100V, and the MC film caps are also 100V rated or better).
Also, from my recollection, on stock DH-500 ground wiring is slightly different than for DH-220 (with L and R hooked up slightly differently).
I'm a bit concerned about one of your previous posts, where John H's instructions had you removing a ground wire (and not necessarily provding another grounding path). Double check with John about ground wiring details--I suspect something is amiss here.
-Bruce
Bruce,
I think you have the solution. I spoke with John Hillig and he said that E13 must connect to ground. In my very literal interpretation of the instructions, the DH-500 appears to not need this connection. It is not shown on the Figure B or Figure C and is not mentioned in the instructions.
I don't understand how the right channel is working although there may be a ground through the RCA plug in some way.
If I have time tonight I will put one of the boards in and see if this is the problem. I will report what I find.
Thanks for the help and hopefully for the solution.
Lindsay
I think you have the solution. I spoke with John Hillig and he said that E13 must connect to ground. In my very literal interpretation of the instructions, the DH-500 appears to not need this connection. It is not shown on the Figure B or Figure C and is not mentioned in the instructions.
I don't understand how the right channel is working although there may be a ground through the RCA plug in some way.
If I have time tonight I will put one of the boards in and see if this is the problem. I will report what I find.
Thanks for the help and hopefully for the solution.
Lindsay
Lindsay,
glad you were able to speak with John H.
Correction: I double checked, and the callout for the "isolated" FE ground is E13 , not E3 as I indicated in all previous posts (E3 is +Rail).
I found some notes from speaking with John H. a while back, about the three-wire grounding setup (use isolated RCA's, E1 to RCA center pin, E2 to isolated RCA shell lug, E13 to chassis gnd lug, and a heavy wire from the chassis lug back to the star ground a junction of main filter caps)
On my DH-220, I measure about 10 ohms between each isolated RCA shell and the chassis ground screw-- my notes show that John H used between 2.2 and 10 ohms for this isolation resistor (depending on board rev).
Hope this resolves LARGE DC offset problem...
-Bruce
glad you were able to speak with John H.
Correction: I double checked, and the callout for the "isolated" FE ground is E13 , not E3 as I indicated in all previous posts (E3 is +Rail).
I found some notes from speaking with John H. a while back, about the three-wire grounding setup (use isolated RCA's, E1 to RCA center pin, E2 to isolated RCA shell lug, E13 to chassis gnd lug, and a heavy wire from the chassis lug back to the star ground a junction of main filter caps)
On my DH-220, I measure about 10 ohms between each isolated RCA shell and the chassis ground screw-- my notes show that John H used between 2.2 and 10 ohms for this isolation resistor (depending on board rev).
Hope this resolves LARGE DC offset problem...
-Bruce
Thank you to Bruce and everyone else that has helped me on this.
It was the ground wire from E13 to the chassis that was the problem. The real question now is why did the right channel work without this ground? At least why did I seem to be able to set the bias correctly?
In any case last night I put in the original board this time connecting E13 to the chassis ground lug between the two RCA connectors and everything is working correctly. I was able to set the bias on both channels (I connected E13 on the right channel as well
). And I was able to get the DC offset within a few hundredths of a volt of zero.
I have not been able to do any extensive listening, but it sounds great so far.
Thanks again for everyone's help. This forum rocks!
Lindsay
It was the ground wire from E13 to the chassis that was the problem. The real question now is why did the right channel work without this ground? At least why did I seem to be able to set the bias correctly?
In any case last night I put in the original board this time connecting E13 to the chassis ground lug between the two RCA connectors and everything is working correctly. I was able to set the bias on both channels (I connected E13 on the right channel as well
). And I was able to get the DC offset within a few hundredths of a volt of zero.I have not been able to do any extensive listening, but it sounds great so far.
Thanks again for everyone's help. This forum rocks!
Lindsay
DH-500
Hi everybody.
Just bought my first DH-500 two days ago, and spent those two days reading all posts regarding DH 200/500 mods.
Here is my question: In many posts mentioned measuring and adjusting the bias. How to measure it (current draw?, DC offset?)? At what points? And how to adjust it?
On my DH-500 I have a board PC-19C, with schematics in
http://www.hafler.com/techsupport/pdf/DH-500_amp_man.pdf
In the manual bias adjustment wasn't detailed enouph for me, so I hoped you could help me.
Now about the mods: my unmodified amp is dead silent, so I cannot see any reasons for the grounding/topology change. And with stock 40,000 uf I have visible drop in voltage when I turn up amp, so to double that value would be probably too much for my house AC breaker?
I need to spend much more time doing comparison with my current Rotel amp to say something about sound quality.
Appreciate your help.
Hi everybody.
Just bought my first DH-500 two days ago, and spent those two days reading all posts regarding DH 200/500 mods.
Here is my question: In many posts mentioned measuring and adjusting the bias. How to measure it (current draw?, DC offset?)? At what points? And how to adjust it?
On my DH-500 I have a board PC-19C, with schematics in
http://www.hafler.com/techsupport/pdf/DH-500_amp_man.pdf
In the manual bias adjustment wasn't detailed enouph for me, so I hoped you could help me.
Now about the mods: my unmodified amp is dead silent, so I cannot see any reasons for the grounding/topology change. And with stock 40,000 uf I have visible drop in voltage when I turn up amp, so to double that value would be probably too much for my house AC breaker?
I need to spend much more time doing comparison with my current Rotel amp to say something about sound quality.
Appreciate your help.
"How to measure it (current draw?, DC offset?)? "
DC is from the red to black posts, adjust P1 (neard board edge) to zero after warmup.
Current draw is 350mA per channel. Adjust one channel at a time. Pull one fuse of the pair for each channel (FL and FR on pictorial), clip your ammeter across the removed fuse, and adjust P2 (middle of board).
DC is from the red to black posts, adjust P1 (neard board edge) to zero after warmup.
Current draw is 350mA per channel. Adjust one channel at a time. Pull one fuse of the pair for each channel (FL and FR on pictorial), clip your ammeter across the removed fuse, and adjust P2 (middle of board).
HaflerDH200 mods
Hi enthusiasts of Hafler, i am pleased to see so much interest in this rather old
design .Like Slowhands, i have been modifying these amps and repairing them
for years and use a version in my 300watt monoblocks. reading all the gen it
seems to me that some would like to go so far with this design that it no longer resembles the hafler at all, others would like to benefit from simple straight forward
improvements of which there are many.Initally i must concur with all the useful
and factual details given by Slowhands, but i think i can contribute a few tips.
As has already been explained the amp can be twitchy and does in some instances
want to ocillate particularly if overdriven or when switching off I have found it crucial
to sort out the input earthing I.E to make both phono i/p sockets isolated then
from the earth pin 2 on the i/p from each board take a wire to an insulated eye and from that a single wire back to the main capacitors centre point this single wire return performs a common mode function and tends to be better at reducing low level hum. it is now necesary to remove R39 the 2.2ohms resistor from the centre of each board thus isolating the i/p and o/p earths, this alone helps the stability of this amp considerably. A further stability measure can be the use of small 22pf
capacitors preferably polystyrene connected from Drain to Gate on each o/p
mosfet which effectively kills any attempts at instability.NB in the uk this amp
had its chassis earthed by three core cable.. isolating all the earths away from the
chassis at the i/p etc stopped any hum loops from associated equipment from
occurring which was quite prevelant from U.S. wired equipment
I hope this is of some use .more to come later after moderation
regards
Hi enthusiasts of Hafler, i am pleased to see so much interest in this rather old
design .Like Slowhands, i have been modifying these amps and repairing them
for years and use a version in my 300watt monoblocks. reading all the gen it
seems to me that some would like to go so far with this design that it no longer resembles the hafler at all, others would like to benefit from simple straight forward
improvements of which there are many.Initally i must concur with all the useful
and factual details given by Slowhands, but i think i can contribute a few tips.
As has already been explained the amp can be twitchy and does in some instances
want to ocillate particularly if overdriven or when switching off I have found it crucial
to sort out the input earthing I.E to make both phono i/p sockets isolated then
from the earth pin 2 on the i/p from each board take a wire to an insulated eye and from that a single wire back to the main capacitors centre point this single wire return performs a common mode function and tends to be better at reducing low level hum. it is now necesary to remove R39 the 2.2ohms resistor from the centre of each board thus isolating the i/p and o/p earths, this alone helps the stability of this amp considerably. A further stability measure can be the use of small 22pf
capacitors preferably polystyrene connected from Drain to Gate on each o/p
mosfet which effectively kills any attempts at instability.NB in the uk this amp
had its chassis earthed by three core cable.. isolating all the earths away from the
chassis at the i/p etc stopped any hum loops from associated equipment from
occurring which was quite prevelant from U.S. wired equipment
I hope this is of some use .more to come later after moderation
regards
hafler DH200 mods continued
My previous comments concerned the stability of this amp,perhaps the most important mod of all....next is a look at the power supplies.On the earlier models
the early stages were not decoupled very well,later 100microfarads[mf] were
added after r28 d11, and r34 d12, labeled c7 and c10 on the board, which was
an improvement..Todays capacitors are much smaller and these caps can now be
increased to at least 220mf in the same size 25mm by 15mm[80-100volts wkg]
Next, some of the amps got very hot and the bias currents seem to vary.I had
put this down to the U.S.A. being at 60hertz and the U.K being 50hertz mains
supply,in any event i did not find any real advantage in running these amps at these
high supply currents especially with such a relatively small reservoir capacitance.
The recommended bias for the later DH220 was 275mA ,now mosfets have a transition of positive tempco to negative tempco around the 80mA region,running
them much above this value will not significantly straighten the mutual conductance
curves, therefore i suggest a compromise in bias of 225mA per side, the early stages
take about 22mA leaving 100mA per pair of mosfets [per side],this gives a little
extra in hand overal giving a better supply stabilization especially with more difficult
speaker loads, then bridging the reservoirs with say 2.2 to 4.7 mfs will help remove
any additional H.F switching junk from the main diode bridge etc. To set the bias
i recommend up a .22ohm resistor
My previous comments concerned the stability of this amp,perhaps the most important mod of all....next is a look at the power supplies.On the earlier models
the early stages were not decoupled very well,later 100microfarads[mf] were
added after r28 d11, and r34 d12, labeled c7 and c10 on the board, which was
an improvement..Todays capacitors are much smaller and these caps can now be
increased to at least 220mf in the same size 25mm by 15mm[80-100volts wkg]
Next, some of the amps got very hot and the bias currents seem to vary.I had
put this down to the U.S.A. being at 60hertz and the U.K being 50hertz mains
supply,in any event i did not find any real advantage in running these amps at these
high supply currents especially with such a relatively small reservoir capacitance.
The recommended bias for the later DH220 was 275mA ,now mosfets have a transition of positive tempco to negative tempco around the 80mA region,running
them much above this value will not significantly straighten the mutual conductance
curves, therefore i suggest a compromise in bias of 225mA per side, the early stages
take about 22mA leaving 100mA per pair of mosfets [per side],this gives a little
extra in hand overal giving a better supply stabilization especially with more difficult
speaker loads, then bridging the reservoirs with say 2.2 to 4.7 mfs will help remove
any additional H.F switching junk from the main diode bridge etc. To set the bias
i recommend up a .22ohm resistor
hafler DH200 mods continued
My previous comments concerned the stability of this amp,perhaps the most important mod of all....next is a look at the power supplies.On the earlier models
the early stages were not decoupled very well,later 100microfarads[mf] were
added after r28 d11, and r34 d12, labeled c7 and c10 on the board, which was
an improvement..Todays capacitors are much smaller and these caps can now be
increased to at least 220mf in the same size 25mm by 15mm[80-100volts wkg]
Next, some of the amps got very hot and the bias currents seem to vary.I had
put this down to the U.S.A. being at 60hertz and the U.K being 50hertz mains
supply,in any event i did not find any real advantage in running these amps at these
high supply currents especially with such a relatively small reservoir capacitance.
The recommended bias for the later DH220 was 275mA ,now mosfets have a transition of positive tempco to negative tempco around the 80mA region,running
them much above this value will not significantly straighten the mutual conductance
curves, therefore i suggest a compromise in bias of 225mA per side, the early stages
take about 22mA leaving 100mA per pair of mosfets [per side],this gives a little
extra in hand overal giving a better supply stabilization especially with more difficult
speaker loads, then bridging the reservoirs with say 2.2 to 4.7 mfs will help remove
any additional H.F switching junk from the main diode bridge etc. To set the bias
i recommend getting a.22 ohm resistor and soldering it across a fuse then replacing
a supply fuse with this [just one] either positive or negative, then measuring with
a digital meter across this resistor a value which should be set to 49-50milivolts
using P1, the variable resistor on the boards.The .22 resistor should be about
1watt rating, repeat proceedure for other side. Don't use the digital meter set in current mode and put in place of the fuse the long wires and high Z can cause
instability . more mods later
regards
My previous comments concerned the stability of this amp,perhaps the most important mod of all....next is a look at the power supplies.On the earlier models
the early stages were not decoupled very well,later 100microfarads[mf] were
added after r28 d11, and r34 d12, labeled c7 and c10 on the board, which was
an improvement..Todays capacitors are much smaller and these caps can now be
increased to at least 220mf in the same size 25mm by 15mm[80-100volts wkg]
Next, some of the amps got very hot and the bias currents seem to vary.I had
put this down to the U.S.A. being at 60hertz and the U.K being 50hertz mains
supply,in any event i did not find any real advantage in running these amps at these
high supply currents especially with such a relatively small reservoir capacitance.
The recommended bias for the later DH220 was 275mA ,now mosfets have a transition of positive tempco to negative tempco around the 80mA region,running
them much above this value will not significantly straighten the mutual conductance
curves, therefore i suggest a compromise in bias of 225mA per side, the early stages
take about 22mA leaving 100mA per pair of mosfets [per side],this gives a little
extra in hand overal giving a better supply stabilization especially with more difficult
speaker loads, then bridging the reservoirs with say 2.2 to 4.7 mfs will help remove
any additional H.F switching junk from the main diode bridge etc. To set the bias
i recommend getting a.22 ohm resistor and soldering it across a fuse then replacing
a supply fuse with this [just one] either positive or negative, then measuring with
a digital meter across this resistor a value which should be set to 49-50milivolts
using P1, the variable resistor on the boards.The .22 resistor should be about
1watt rating, repeat proceedure for other side. Don't use the digital meter set in current mode and put in place of the fuse the long wires and high Z can cause
instability . more mods later
regards