HI rhr9
I think the top and the bottom of the transformers mounted copper foil. Otherwise, I am working as a service engineer for a French company from Montpellier HORIBA ABX (Hematology and Biochemistry analyzers) and almost every day we have excellent contacts and cooperation for almost 20 years. Otherwise I visiting them because of training and by little I learned French for ease of reference.
"only if there has been a short circuit on the unit"
A DH500 can drive 80V peaks on program material into 4Ω on each channel, that's a 2Ω load on the supply. 80V/2Ω=40A peak current. This current has to be provided by the filter caps through the 16ga wire (the transformer can only provide a small fraction of this without the caps).
I am assuming that this wire on the DH500 broke due to vibration (amp was in a touring sound system). The DH500 was about five years old when it happened. I picked up a big roll of copper strap for about $1 (for a pound).
"(power switch most important contact)"
I added inrush limiters to all my Hafler amplifiers, no more power switch problems.
A DH500 can drive 80V peaks on program material into 4Ω on each channel, that's a 2Ω load on the supply. 80V/2Ω=40A peak current. This current has to be provided by the filter caps through the 16ga wire (the transformer can only provide a small fraction of this without the caps).
I am assuming that this wire on the DH500 broke due to vibration (amp was in a touring sound system). The DH500 was about five years old when it happened. I picked up a big roll of copper strap for about $1 (for a pound).
"(power switch most important contact)"
I added inrush limiters to all my Hafler amplifiers, no more power switch problems.
When turning on HAFLER DH-200 peak is around 3.3 A per each chanel I measured with a Fluke 179 that is really fast, which means approximately 6.6A-on to both channels. As for the maximum load I doubt that I could test it and get to the true value. As far as DH-500 but I would say that you have no luck these were the circumstances. About current limiter I think is good idea, thinking to put it on my device anyway.
Input Caps & C12 & C14
I have been watching this thread for a long time now. Great stuff but when I read the points on the input cap I still have an issue that I am sure I am just not getting. I see ranges suggested from 2uf to 10 uf.
I have the Hafler DH-220 specifically. It came with a 2uf input cap. If you go up to say a 8uf cap would you not be losing bass in that circuit?
I also think that the C12 and C14 caps which are 100uf at 100v would be best at no less than 330uf at 100v as this seems like an easy application for the best benefit.
Any thoughts?
Thx in advance.
I have been watching this thread for a long time now. Great stuff but when I read the points on the input cap I still have an issue that I am sure I am just not getting. I see ranges suggested from 2uf to 10 uf.
I have the Hafler DH-220 specifically. It came with a 2uf input cap. If you go up to say a 8uf cap would you not be losing bass in that circuit?
I also think that the C12 and C14 caps which are 100uf at 100v would be best at no less than 330uf at 100v as this seems like an easy application for the best benefit.
Any thoughts?
Thx in advance.
Input cap
DH-200 had 10uf and the low frequency roll off point was lower than the one from the feedback which is not preferred... Given the relatively high roll off of the feedback network It is probably better to lower the value of input cap to 2uf or so , Or increase the feedback cap to keep a high value input cap. More bass when not properly controlled is not better to my taste...
Fab
You would be increasing bass.
DH-200 had 10uf and the low frequency roll off point was lower than the one from the feedback which is not preferred... Given the relatively high roll off of the feedback network It is probably better to lower the value of input cap to 2uf or so , Or increase the feedback cap to keep a high value input cap. More bass when not properly controlled is not better to my taste...
Fab
So on my DH-220 the feedback cap is a 220uf non-polar cap and not the 470uf I see from time to time mentioned in this thread and which is also stated in my manual from Hafler. The unit is a factory built unit and when I got it you could instantly tell it WAS NOT touched prior to me opening it up.
So with all that said if I understand you correctly if I was to put one of my Mundorf gold caps at 8.2uf in for the 2uf input cap it would be better if my feedback cap was 470uf correct?
The bass cut-off frequency of a Hafler 200 is straight forward. If you look at the schematic, you see a 6db/octave high pass filter with the values:
R3 = 22K
C1 = 10K
Given f = 1/(2 phi * R * C), we get the -3db corner of the filter is .7 Hz
On the Hafler 220 the values have been changed some and the DC offset added, but it is still a classic 6dB/octave RC filter. So now we have:
R3 = 47K
C1 = 2uF
So for the Hafler 220 the cut off frequency has been moved up to 1.7Hz.
There is no reason in to move to 10uF for the input on a 220. Besides, the bigger the cap, the worse the high end...
If you are after better bass slam, it is better to add to the PS caps. Here the Haflers had 10,000uF. With today's better caps, you can easily triple that.
D.
R3 = 22K
C1 = 10K
Given f = 1/(2 phi * R * C), we get the -3db corner of the filter is .7 Hz
On the Hafler 220 the values have been changed some and the DC offset added, but it is still a classic 6dB/octave RC filter. So now we have:
R3 = 47K
C1 = 2uF
So for the Hafler 220 the cut off frequency has been moved up to 1.7Hz.
There is no reason in to move to 10uF for the input on a 220. Besides, the bigger the cap, the worse the high end...
If you are after better bass slam, it is better to add to the PS caps. Here the Haflers had 10,000uF. With today's better caps, you can easily triple that.
D.
Input cap
You are correct. However, the 8.2uf would be still a little high... Like anything everything is compromise ...
Fab
You are correct. However, the 8.2uf would be still a little high... Like anything everything is compromise ...
Fab
This all makes perfect sense now as these are the mods I made so far that brought me to my original question a couple of posts back.
1. I changed out the 2 original 10K PS caps with Elna Audio grade 30K/side caps @ 100v.
2. Ditched the old bridge and went with fast diodes on sinks all jumped with film caps and drain resistors.
3. Replaced the input jacks with isolated gold plated ones but otherwise left the grounding scheme original since I had no hum issues previously.
Tried this out in a listening test for about 2 weeks and for sure the amp sounded better than original.
Than did the following:
1. Changed out C12 & C14 with 330uf @ 100v muse caps, KZ series.
2. Changed out the 2uf with an 8.2uf.
3. Changed all the PS wiring to 12 gauge.
4. Jumped eyelet 6 to 8 and ran out the speaker connection from eyelet 5 to the fuse than to the positive speaker jack.
5. All speaker wire went to 12 gauge as well.
6. New speaker jacks went in and dumped the old ones.
7. Placed an inrush limiter in line with the 110v prior to the switch.
8. Ditched the temp switches on the heat sinks and simplified the 110v power connection.
9. Replaced the power cord to a 3 wire unit.
10. Rearranged the grounding scheme to single point than dropped signal ground to chassis via a 10 ohm resistor.
Everything from previous got better except the high end cut off so that is why Vilfort's point of view makes sense.
I a going to go back to a 2uf M-cap and call it a day.
Now I did all of the above on only one channel.
The other channel I have on the bench and I am thinking that I will do Vilfort's Version 6 tube input mosfet out put and see what I get.
I will take pics and up load shortly.
Thx
Guys, I'm sad to report, but diyAudio member Dick West (Richard Westenskow) has passed away at 82.
diyAudio - View Profile: Dick West
He will be missed by the community and I will miss him deeply. He knew everything that there is to know about Hafler amps. I learned a great deal from him! He also was the guy to go to if you needed your MOSFETs tested and matched. He had a rare Hafler MOSFET Checker/Matcher. He was a very kind person.
Rest in peace.
Rockford Register Star
Obituary Notices: Week of Feb. 27-March 5, 2013 | The Rock River Times
diyAudio - View Profile: Dick West
He will be missed by the community and I will miss him deeply. He knew everything that there is to know about Hafler amps. I learned a great deal from him! He also was the guy to go to if you needed your MOSFETs tested and matched. He had a rare Hafler MOSFET Checker/Matcher. He was a very kind person.
Rest in peace.
Rockford Register Star
Obituary Notices: Week of Feb. 27-March 5, 2013 | The Rock River Times
Last edited:
RIP Dick
I sold some Hafler DH120 boards to Dick for one of his projects. We had a nice email exchange where he gave me a lot of advice on my Hafler mod projects. He spent a lot of time thinking about my problems so I really appreciated his patience and insight.
The Forum will miss him. RIP, Dick.
I sold some Hafler DH120 boards to Dick for one of his projects. We had a nice email exchange where he gave me a lot of advice on my Hafler mod projects. He spent a lot of time thinking about my problems so I really appreciated his patience and insight.
The Forum will miss him. RIP, Dick.
It is a copper strap that runs all the way around the transformer core and windings. It eliminates some of the vibration of early Hafler DH-200 power transformer laminations, and greatly reduces EMI in any transformer. The original name is "Faraday Shield." I will leave it to Google to explain in more detail, and with pictures or drawings.
Hi Fab,
theorically, is there a limit on the size of the feedback cap in terms of stability?
If my calculations are correct (DH200), the -3dB of the input cap is ~0.7Hz while the feedback rolloff is ~3.4Hz.
Reducing the input cap to 4.7uF gives ~1.5Hz, increasing the feedback cap to 1.000uF gives ~1.6Hz. Is it worth to increase further the feedback cap or this could be considered a good solution?
Ciao
theorically, is there a limit on the size of the feedback cap in terms of stability?
If my calculations are correct (DH200), the -3dB of the input cap is ~0.7Hz while the feedback rolloff is ~3.4Hz.
Reducing the input cap to 4.7uF gives ~1.5Hz, increasing the feedback cap to 1.000uF gives ~1.6Hz. Is it worth to increase further the feedback cap or this could be considered a good solution?
Ciao
The power supply caps form a pole with the load, the feedback pole must be higher than the power supply pole, and the input pole must be higher than the feedback pole.
10,000µF driving two 4Ω loads is about 8hz.
Changing the input to a 12dB filter with a Q=1.0 really helps, and raising the feedback pole to be between the input and power pole helps too.
For best transient response and settling performance the input pole must predominate, followed by the feedback pole, all determined by the power supply pole.
Look at the input of the NAD 3020 to see an example of a Q=1 filter.
10,000µF driving two 4Ω loads is about 8hz.
Changing the input to a 12dB filter with a Q=1.0 really helps, and raising the feedback pole to be between the input and power pole helps too.
For best transient response and settling performance the input pole must predominate, followed by the feedback pole, all determined by the power supply pole.
Look at the input of the NAD 3020 to see an example of a Q=1 filter.
Hi pidigi
your selected values are a good start. Personally I would ( bacause I haven't work on a DH-200 for years but made plenty of other amps and most with lateral MOSFET ):
1) increase power supply caps from 10kuf to 27-30kuf;
2) feedback cap to 1000uf (good quality)
3) input cap of 2.2-3.3uf (non polar)
You could also experiment to change the associated resistors values but you need to know what you are doing before any attempt...
Good luck
Fab
"Any direct input you can supply? Schematic of parts change? Thx "
Sign in (free)
http://http://www.hifiengine.com/library/nad/3020.shtml
Add 1µF between input jack and board, 1µF (C1) with (R3) 22K to ground, and 5.6K between the feedback node (junction of R20, R21) and the node between the added input cap and C1 = 14.35hz, Q=1.
Change C5 to 150µF~160µF with the stock 10,000µF filter caps, may be a larger value as supply C is increased.
Sign in (free)
http://http://www.hifiengine.com/library/nad/3020.shtml
Add 1µF between input jack and board, 1µF (C1) with (R3) 22K to ground, and 5.6K between the feedback node (junction of R20, R21) and the node between the added input cap and C1 = 14.35hz, Q=1.
Change C5 to 150µF~160µF with the stock 10,000µF filter caps, may be a larger value as supply C is increased.
Last edited: