Need I say more?
Not as smooth as the McIntosh (duh!) but VASTLY improved. A bit brighter than the Mc (I’m told by the owner of the Mc that emulation of tube sound was an objective in Mc SS design). And I know my VOTT’s horns are going to be bright at residential listening distances, thus some horn fatigue after long listening sessions. If the larger main caps will improve the bass as most everyone agrees in this thread, I think I’ll be pretty close to audio nirvana for this setup.
Thank you, Fab for sticking with me. I guess the folks who have bypassed the circuit in question and had good results might have made other mods that I have not.
Other than the increased main caps (if I can find some. Wincing at the prospect of spending $50-75 for the Nichicon or Kemet units each as opposed to waiting for the Cornell Dubilier $27 ones) I think that will be the extent of mods. I will replace those WIMA caps on the mosfets with ceramic ones when they get here.
Thanks again, Fab!
Jeff
Hi Jeff
Thanks.
I am glad I could somehow help.😎
Indeed I prefer new main power supply caps over reformed caps…but maybe it is only me…also look for low ESR value.
Also, consider changing input cap C1 for a Polypropylene type. 10uF is overkill as a value And would be quite big physically…I successfully used 3.9uF without any loss of bass for this similar input stage impedance. In fact, it is more the quality of the specific cap itself that could “improve” the bass more than the lowest possible value. Thus, do not hesitate to pay more than 20$ on this cap And you could rewarded in the overall tonal balance…
Replacing the feedback cap C5 with a good one is to be considered too. Maybe Muse Nichicon (but not for input stage) or Elna Silmic…
Good luck in your journey And have fun in the process.🙂
Fab
Thanks.
I am glad I could somehow help.😎
Indeed I prefer new main power supply caps over reformed caps…but maybe it is only me…also look for low ESR value.
Also, consider changing input cap C1 for a Polypropylene type. 10uF is overkill as a value And would be quite big physically…I successfully used 3.9uF without any loss of bass for this similar input stage impedance. In fact, it is more the quality of the specific cap itself that could “improve” the bass more than the lowest possible value. Thus, do not hesitate to pay more than 20$ on this cap And you could rewarded in the overall tonal balance…
Replacing the feedback cap C5 with a good one is to be considered too. Maybe Muse Nichicon (but not for input stage) or Elna Silmic…
Good luck in your journey And have fun in the process.🙂
Fab
Last edited:
Update:
Appropriate suggested parts are hard to come by but should be on their way soon.
Concerned about possible phasing issues with my subs I ran some tone tests today and besides finding and correcting a phasing error, I also am sensing an oscillation above 300hz or so. It is subtle but there. Haven’t dragged the o’scope out of the basement bench yet but it sounds like about 4 to 5 hz both channels individually and together. Sounds like a subtle vibrato.
Is this the oscillation mentioned here on occasion, and if so will the capacitor change’s recommended above address it or should I look elsewhere?
Thanks,
Jeff
Appropriate suggested parts are hard to come by but should be on their way soon.
Concerned about possible phasing issues with my subs I ran some tone tests today and besides finding and correcting a phasing error, I also am sensing an oscillation above 300hz or so. It is subtle but there. Haven’t dragged the o’scope out of the basement bench yet but it sounds like about 4 to 5 hz both channels individually and together. Sounds like a subtle vibrato.
Is this the oscillation mentioned here on occasion, and if so will the capacitor change’s recommended above address it or should I look elsewhere?
Thanks,
Jeff
Hi guys! I am rebuilding a DH200 presently, have made a huge pile of notes thanks to many of you, and printed out all the build and circuit diagrams, its a pretty easy amp to study. Bear in mind I am not a trained tech for electronics, Industrial Design is my background and have self taught all my electronics knowledge over the past 30 odd years.
There have been quite a few interesting mess-ups with this particular unit, one of which was it was being run on 120v step down transformer when I got it (Aussie 240v here) yet it has the international transformer and it was configured for 240! So how do you ask is this possible? Well after sitting staring at it for an hour and going over the drawings, I suddenly realised the switch had been wired back to front. Someone replaced the switch, and wired the active directly to pin 2, with pin 1 going to the jumper link which is connected to the Tx windings BR & BK/W which I guess doubles voltage hence only needing 120vac input?!? The other side which runs through the TB's was connected to pin 3 when it should have been on pin 1.
Anyway reversing the switch wiring had it running on 240v again and things measured ok, dc offset was not too high, 10mv and 40mv, and the amp sounded pretty good.
I have both heatsink modules off in order to clean the horrible black paint that was sprayed all over the amp to "freshen" it up 😀 and I am rebuilding the amp boards, new caps, replacing some of the resistors that have been listed as needing to be spot on in order to keep the Q pairs happily balanced etc etc.
So here is the question I have for all the experts in here. The output transistors have base stop resistors of 470ohms, so these have been upped from the original circuits 220ohms. The 390pf mica cap C13 is present. But the 680pf mica cap C18, which appears to go from the + side? of the transistor is missing, instead there is a 47pf cap attached at the same point as the base stop resistor, going to the ground terminal of the tag strip (neither of which were bolted to the heatsink so not making proper ground anyway). This attachment point is different to every photo I have scoured online, and different to the actual circuit diagram showing where C18 should be connected. Please see attached photo, I have removed C13 and unsoldered the orange wire so things are more visible.
Now I read in my notes someone said the output transistors had power bypassing caps on the DH220, while the DH200 does not. They suggest adding these 0.1uf caps to chassis ground, but where do they attach to on the transistors? If someone could suggest how I should proceed here it would be much appreciated. I have replaced C13 with a 390pf silver mica, am not sure on what the 47pf cap is doing in the pic, and where I should attach the 0.1uf caps (getting rid of the 47pf at the same time I guess).
There have been quite a few interesting mess-ups with this particular unit, one of which was it was being run on 120v step down transformer when I got it (Aussie 240v here) yet it has the international transformer and it was configured for 240! So how do you ask is this possible? Well after sitting staring at it for an hour and going over the drawings, I suddenly realised the switch had been wired back to front. Someone replaced the switch, and wired the active directly to pin 2, with pin 1 going to the jumper link which is connected to the Tx windings BR & BK/W which I guess doubles voltage hence only needing 120vac input?!? The other side which runs through the TB's was connected to pin 3 when it should have been on pin 1.
Anyway reversing the switch wiring had it running on 240v again and things measured ok, dc offset was not too high, 10mv and 40mv, and the amp sounded pretty good.
I have both heatsink modules off in order to clean the horrible black paint that was sprayed all over the amp to "freshen" it up 😀 and I am rebuilding the amp boards, new caps, replacing some of the resistors that have been listed as needing to be spot on in order to keep the Q pairs happily balanced etc etc.
So here is the question I have for all the experts in here. The output transistors have base stop resistors of 470ohms, so these have been upped from the original circuits 220ohms. The 390pf mica cap C13 is present. But the 680pf mica cap C18, which appears to go from the + side? of the transistor is missing, instead there is a 47pf cap attached at the same point as the base stop resistor, going to the ground terminal of the tag strip (neither of which were bolted to the heatsink so not making proper ground anyway). This attachment point is different to every photo I have scoured online, and different to the actual circuit diagram showing where C18 should be connected. Please see attached photo, I have removed C13 and unsoldered the orange wire so things are more visible.
Now I read in my notes someone said the output transistors had power bypassing caps on the DH220, while the DH200 does not. They suggest adding these 0.1uf caps to chassis ground, but where do they attach to on the transistors? If someone could suggest how I should proceed here it would be much appreciated. I have replaced C13 with a 390pf silver mica, am not sure on what the 47pf cap is doing in the pic, and where I should attach the 0.1uf caps (getting rid of the 47pf at the same time I guess).
Attachments
Hi Sonic Art,
Just one comment on your DC offsets. When the diff pairs are properly matched I often get DC offsets less than 10 mV. Your channel with a 40 mV offset has poor matches. Distortion also improves with good matches.
For your gate stopper resistors, I tend to leave them at original values unless I see instability. Your amp may have been worked on and oscillation "cured" by empirical means. Not always the right way.
-Chris
Just one comment on your DC offsets. When the diff pairs are properly matched I often get DC offsets less than 10 mV. Your channel with a 40 mV offset has poor matches. Distortion also improves with good matches.
For your gate stopper resistors, I tend to leave them at original values unless I see instability. Your amp may have been worked on and oscillation "cured" by empirical means. Not always the right way.
-Chris
Hi Jeff
Has this problem been fixed now.
if not, this seems like a “motorboating“ issue. You can search the web for explanations. These may be difficult to fix in a new amp but this one has been design proven for years…
From my humble experience, in these situations it can possibly be fixed by securing again all your grounding connections (even RCA plug ground sometimes not doing a very good connection) or revisit your grounding scheme….
Fab
Check the DH-220 schematic available for Hafler's (Rockford) website in the manual. It's all there. Familiarize yourself with Mosfet terminology and function. For the DH- 220, N-Channel Gate resistors are 470ohm and P-Channel are 220ohm. The .1uf should go from the drain to ground at the second output fet on the N-channel (where the 47pf is now) and from Drain to ground on the second P-Channel fet per the DH-220 schematic. That terminal strip is connected to the ground where it bolts to the heatsink and that center solder point. The 680pF should go across the second N-channel Gate and Source. Best of Luck!
For those who are curious, the different values of gate resistors help match the slew rates of the P and N devices. This was demonstrated by Nat Semi (now TI) in an application note on their now discontinued BJT/MOSFET driver, the LM4702.
Yes, exactly. The reason for this are the very different gate to source capacitances of N and P channel devices.
Best regards!
Best regards!
TI made some great app notes for both audio and instrumentation. I downloaded this one a few times in the past. Thanks for the link Jack.
One valuable point made in this app note is the need for a real driver stage when driving Mosfets. Most "designers" don't feel the need for one. Gate charge is a thing too.
One valuable point made in this app note is the need for a real driver stage when driving Mosfets. Most "designers" don't feel the need for one. Gate charge is a thing too.
That's some good information! Makes me think those gate resistor values should be looked at when substituting exicons since the input capacitance differs from the original Hitachi. Another good article on MOSFET is: https://toshiba.semicon-storage.com/info/docget.jsp?did=13415 .
The DH220 also appears to be directly matching input capacitance between N and P with the 680uf across the N channel GS. Looking at the Hitachi datasheets, the input capacitance between P and N differs by 300pF (at 10v 1MHz). Possibly the value was empirically derived with a scope over different frequencies and voltage. Again, this makes think the exicon substitution might require revisiting the value of that GS cap.....
Hi pwayland,
Yes, absolutely!
Even with BJTs, you should be looking at the compensation if you change output types / brands (normally close enough with the same number).
Yes, absolutely!
Even with BJTs, you should be looking at the compensation if you change output types / brands (normally close enough with the same number).
Thanks I do have all the diagrams for both 200 and 220, I am trying to get my head around the pin connections. Can we use my photo as reference for orientation?
Looking at the photo, the centre pin, which is connected to the case of the transistor is the source yes? then the lower pin in the photo is the gate (pin 1) and the upper pin (2) is the drain.
So the 47pf is connected to the wrong pin then? as it appears to be connected to the lower pin (1) which is the gate?
For amplifiers that use mosfets as source followers (most often the case), the gate-drain capacitance is likely a bigger problem than the gate-source capacitance. Gate to drain has a bigger voltage swing/differential than gate-source, so the non-linear capacitance variance vs voltage swing is a bigger problem. Remember, in source follower mode, the input capacitance is bootstrapped, albeit only to limited degree; since lateral mosfets have relatively low Gm!
and now in English? 😀
so are you saying the 47pf cap connected from the gate to ground is ok even though it's not part of the original design? this amp has no cap from drain to ground, only what I described in my first post, the original circuits show 680pf or 0.1uf going from drain to ground with the smaller pf cap across gate and source right?
so are you saying the 47pf cap connected from the gate to ground is ok even though it's not part of the original design? this amp has no cap from drain to ground, only what I described in my first post, the original circuits show 680pf or 0.1uf going from drain to ground with the smaller pf cap across gate and source right?
anyone? 121 pages of knowledgeable ppl posting but no answers to some simple questions from someone just wanting clarification that what they are interpreting from diagrams is correct?
I can read the circuits, I can see where things "should" go, but I like to be sure that my interpretation is correct, I just want to know that I'm seeing the pin numbers correctly, identifying their task correctly, anyone?
I can read the circuits, I can see where things "should" go, but I like to be sure that my interpretation is correct, I just want to know that I'm seeing the pin numbers correctly, identifying their task correctly, anyone?
Search online for the datasheet of the lateral mosfets you have in your amp and it will have a picture of the pinout. Check both The N type and P type
Hope this helps
- Dan
Hope this helps
- Dan
Yes I got that diagram already, all I was asking was for confirmation that what I was saying (from what I was understanding reading and looking at the amp in question) was correct before I go changing things around from how they came.
I'm traveling & don't have access to my notes, but here are a couple thoughts.
1. Bob Cordell has revised the Hafler amps and only keeps the FETs, Transformer & thermostats, as I recall. A Canadian fellow did the PCBs including for the DH-200, 220, 280 and others. The raw PCBs and comprehensive notes are available on eBay. More info is available on Bob's site.
2. To keep the transformer healthy and the switch working long term, use a big MOV (Mouser, Digikey, etc.) from Antherm or others. The size will depend on the amp. I'd start with about 10 Ohms and look for a 1 or 2 Volt drop between the AC line and amp after the MOV heats up. The Antherm site has sizing info. An alternative is Antek, who make / import toroidal transformers and have an inexpensive relay switched resistors unit that comes in a metal case.
3. I'd also check the FETs for matching. The badly printed numbers on the TO-3 cans should be the same for each bank of FETs. Otherwise, Exicon still has some matched TO-220 Lateral FETs left. Search for Nelson Pass's article on FET matching. While you are at it, replace all the poor quality gate resistors with metal film, 1% types for lower noise and greater stability. Use low tempco types.
1. Bob Cordell has revised the Hafler amps and only keeps the FETs, Transformer & thermostats, as I recall. A Canadian fellow did the PCBs including for the DH-200, 220, 280 and others. The raw PCBs and comprehensive notes are available on eBay. More info is available on Bob's site.
2. To keep the transformer healthy and the switch working long term, use a big MOV (Mouser, Digikey, etc.) from Antherm or others. The size will depend on the amp. I'd start with about 10 Ohms and look for a 1 or 2 Volt drop between the AC line and amp after the MOV heats up. The Antherm site has sizing info. An alternative is Antek, who make / import toroidal transformers and have an inexpensive relay switched resistors unit that comes in a metal case.
3. I'd also check the FETs for matching. The badly printed numbers on the TO-3 cans should be the same for each bank of FETs. Otherwise, Exicon still has some matched TO-220 Lateral FETs left. Search for Nelson Pass's article on FET matching. While you are at it, replace all the poor quality gate resistors with metal film, 1% types for lower noise and greater stability. Use low tempco types.