[12796]

Looking forward to it! Good luck with the move!

 

[PinkFloyd]

Quote:

Originally Posted by biovizier Good luck with the move!

Haven't got a clue where I'm going yet Nick but I'm going somewhere for sure. I've made a big change to my life so may as well change the location whilst I'm at it....... It'll probably be somewhere as far away from "civilisation" as possible.

Mike.

 

[spendorspain]

Hi head-fiers,

Regarding the input and output caps, I believe that they are the limiting factor in the WNA hp amp.
When I removed input caps the sound improved vastly, but I had to use output caps because the bad dc-offset behaviour of the LM6171. I was sure that the sound quality of Black Gate caps I used in C4 was beyond reproach, but a couple of simple experiments proved this was wrong. I began to believe that the sound was perhaps too nice and “uniform”. When I used other amp to drive my Grado HP-2, the recordings seemed more different and distinct one from another, so I suspect something was putting a little veil over them, a very charming and rounding veil, but a veil. Of course the amp did sound very well,
but perhaps it wasn’t telling the whole truth of what there is in the recording. So I reconnect my cans to the Copland CTA501 valve power amp binding posts via an high quality attenuator. The result was a more live sound and the restored feeling that all recordings are clearly different, as if the acoustic personality of each were more clearly presented, even sounding coarser in some cases. Next, I made an experiment that I recommend to everybody that has removed the 4.7uF C1 input caps (a high quality Evox polypropylene film caps) and has them, unused, in the spare parts box. I soldered them across the C4 legs under the PCB, as a bypass cap. It seemed a too big cap for bypassing, but C4 has also a big value (1000uF in my amp), and I had at hand only the spare C1 to experiment, so I went ahead. The result was an improved sound, specially in extreme highs and in what could be called as the “aliveness” of sound. This impression is confirmed by nearly a month of listening to the “bypassed” amp. This disappointed me a little, because the Black Gates are touted as possibly the best electrolytics and they are very very expensive… It seems that they didn’t need a “band-aid” (the film bypass cap) to improve its sound, but IMHO this is not the case, and perhaps this is ever more important with the stock Nitai output caps. I’d like to know more opinions about this.
As a result, I’m decided to build another WNA hp amp, this time “cap-less” and as simple as I can. No caps in signal path, no need for rail splitter (I’ll use a WNA Cascode psu to supply the +/0/- dc volts), no zobel network and the best parts I could find (there are only 23 per channel). I have already a pair of AD8610 chips (with extremely low dc-offset specs) mounted on Browndog adaptors and I’m going to order the rest of parts from several suppliers (WNA, Mouser, Percy…). I have no problem with resistors, capacitors and output transistors, but I have no idea of what are the exact references of the diodes D1-D2 and the constant current diodes I1 to I4. I’d be very grateful if somebody could tell me these references (type, maker, exact values). When finished, I hope to report about the differences (to my ears and with my system, of course) between the cap-less AD8610 WNA amp and the standard LM6171 plus output caps (note that even with the 100k trimpot and 10M resistor trick proposed by Pink Floyd to eliminate DC offset, it is necessary to use input caps). And I must add that the WNA hp amp I’m listening now, caps and all, is a very enjoyable and very good sounding little amp. Perhaps it’s just me and a case of “audiophilia nerviosa”…

Jose

 

[PinkFloyd]

Hi Jose the current regulating diodes are 3.5mA and 5.6mA made by Semitec. You can source them at Rapid for 55p each

The in / out caps certainly do throw a slight rose tinted veil over the music and it is well worth bypassing them (preferably with a zero ohm link!) try a 1uF poyprop and say a 200nF film cap if you're intent on keeping the 6171 in the circuit.

From memory, the AD8610 wasn't anything spectacular but I tried it in a capped MKl so it'll probably sing a bit better with them removed. There is no need to have in / out caps if you're using a FET opamp so you must choose which FET opamp sounds the best to your ears. I'm currently enjoying the AD-8605 which sounds honest to the event and I like it. Nick has sent me a couple of AD8067 to try so I'll report back on them.

Black gates are great in certain applications but I'm not sure the signal path is one of them...... caps in the signal path are always a compromise and I'm not conviced that the LM6171 (with caps) is anywhere near as good as a 8065 (without caps) for example.

If you give me a couple of days I've got a few ideas for you to try out........ I must get off to bed now else I'll be up all night hoovering and painting.

Mike.

 

[bludragon]

Fitted the trim pots last night so I could remove the output caps and leave the LM6171 in place. Everything worked perfectly - adjusted the offset, waited till it warmed up, and then adjusted it again. After a couple of hours of it working away with no lid, decided to unplug it.

Now when I plug in the supplied power supply, it drops from 24V down to 3V. I tried a different 12V supply, and this works OK. I put a 1K Ohm resistor across the 24V supply, and its voltage didn't drop at all. So it seems more like something has gone wrong in the WNA which makes the power input look like a short at higher voltages??

Will be investigating more fully tonight, but any suggestions as to what to check, or what might be wrong will be very much appreciated.

Thanks.

 

[StevieDvd]

Have you got a ground wire on the volume pot? My WNA had a hiccup reently and after a long autopsy it was the wire that was part of the cause. This led Mike (Pinkfloyd) to have his doubts about the LM6171 as he has mentioned earlier in this thread.

If you do have the wire there in case of 'hum' it's a quick check to do to eliminate it as a cause.

Steve

 

[bludragon]

No ground wire. The mods I do have are:
Zobel network
incandescant lamp replaced with zero ohm link
some of the caps bypassed smaller values.

and now zero ohm link instead of output caps, and trip pots to adjust offset.

I think its more likely something has died due to a possible short or other electrical zap when I was messing around with the top off, as it was working fine for a couple of hours.

 

[1UP]

Jose, those bypass caps are huge! How tall is your case?

Which BGs were you using in which position, exactly?

 

[PinkFloyd]

Quote:

Originally Posted by StevieDvd Have you got a ground wire on the volume pot? My WNA had a hiccup reently and after a long autopsy it was the wire that was part of the cause. This led Mike (Pinkfloyd) to have his doubts about the LM6171 as he has mentioned earlier in this thread. If you do have the wire there in case of 'hum' it's a quick check to do to eliminate it as a cause. Steve

Not so much "doubts" about the LM6171 Steve just "puzzlement" when the **** hits the fan at times. David White reckons it's easier to design an amp than it is to pinpoint a fault..... seems to me there are no "obvious" places to look when a fault or weird noise develops it's just a case of process of elimination and sods law dictates that the last part you replace is the bloody part that has failed or is misbehaving..... It's never the "first" part you check that's faulty.... oh no, never works out like that!

Quote:

Originally Posted by bludragon Will be investigating more fully tonight, but any suggestions as to what to check, or what might be wrong will be very much appreciated.

Your finger is the best (and cheapest) tool to use to begin with. Possible that one or more of the transistors has packed in... if they are working they'll be hot to the touch (the ones in the rail splitter will be cool so that's ok) If you probe one (ampside) with your finger and it's cold then you can assume that it's not working....... if cold then replace and see what happens.

I've e-mailed Dr. David with your symptoms...... maybe he can throw some light on things

Mike.

EDIT: Are you using one of those "powerpak" PSU's by any chance? Just re-read your thread and you say the amp works ok with a different PSU?? what PSU does it work OK with and which one doesn't it work with?

 

[bludragon]

Thanks Mike, I've just spend the last couple of hours with it, and have come to the (possibly incorrect) conclusion, that the fault is in the rail splitter section. With no load on it, it seems to output the correct voltage, but as soon as I attach either one of the amp channels, it drops down to +-1 Volt.

I happen to have a +-12V regulated psu, and with that powering both channels (bypassing the rail splitter) everything seems ok.

I'm about to try plugging some (cheap) headphones in to see what happens...

EDIT: ok, seems to work when bypassing the railsplitter. As its screwed back to gether, I'm gonna leave it at that for tonight, and troubleshoot the railsplitter later.

 

[bludragon]

Just seen your edit - The problem occurs when using the 24V powerpack that comes with the kit - so far I've been having fairly decent results with it. Last night I tried with a different 12V powerpack, and things seemed to work ok. However, there doesn't seem to be anything wrong with the 24V powerpack. At the moment it would seem the railsplitter goes into short circuit mode when asked to supply more than a certain amount of current

 

[PinkFloyd]

Quote:

Originally Posted by bludragon Just seen your edit - The problem occurs when using the 24V powerpack that comes with the kit - so far I've been having fairly decent results with it. Last night I tried with a different 12V powerpack, and things seemed to work ok. However, there doesn't seem to be anything wrong with the 24V powerpack. At the moment it would seem the railsplitter goes into short circuit mode when asked to supply more than a certain amount of current

Hi Bludragon,

Had a word with David re: the power supply etc. the short of it is that certain power supply units can be problematic even with slight changes such as different opamps / layouts etc. I know, first hand, that the Mascot PSU is a no no with the WNA (ok for about 30 minutes and then it trips out) The fact that the "powerpak" was ok and now isn't makes this a very tricky one to pinpoint but, basically, these things happen. Remember the LM6171 is pretty cranky / insanely fast and any changes to the layout can cause it to misbehave.

I know this sounds a bit vague but things like this sometime happen... even the way you wire the trimpots is going to play a part..... wiring needs to be kept as short as possible when using the LM6171..... try shifting one of the ampside opamps to the rail splitter and the one in the rail splitter ampside..... you'd be amazed just how often this does the trick

Mike.

 

[FritzS]

My questions to David:
Have you test some other output transistors - BD138/139 vs. Toshiba's SC2238 / 2SA968 or 2SA1837 / 2SC4793?

Toshiba's 2SC2238 / 2SA968 have a higher ft - about 100 MHz

What costs the headphone protection too - only PCB / complete kit - must I order only one or two for stereo?

The answer:
Quote:

Hi Friedrich, I must admit that I haven't got around to trying any output devices other than the BD139/BD140, but its possible that higher gain and/or ft devices might make a small difference. The headphone protection unit costs £25 as a kit but at present I don't offer the pcb separately. Regards, David

 

[spendorspain]

Hi, 1Up,

The red Black Gates (1000uF/25V Nx bipolar type) are in C4 position, the DC-blocking cap in series with the output, hence its great influence in the final sound of WNA amp. As you can see in the picture, I had to solder the IC bypass small cap C2 in PCB underside to make room for C4. My case is very tall simply because I have removed its lid

. When I decided to bypass C4 (with the 4.7uF polypropylene I had at hand after remove it from C1 position), I thougth it would be only a brief test, so I secured the PCB to the open case, after soldering the bypasses, with small blobs of Blu-Tack. It didn't look pretty, but it worked. However, because of the big and unexpected improvement in sound, I have listened to the amp without lid for several weeks. With a more careful layout and bending the cap leads I think it would be possible to install these big film bypasses in a space of just 2-3 cm under the PCB, so a very big case would not be necessary, just taller stand-offs.

I've already installed the AD8610 in IC1 position, and I've made a temporary wire arrangement to compare the sound with and without output caps, just to hear the effect of BG in C4. To hear the amp with caps, I connect the cable to headphone socket to the P2 pin header (the normal way); to hear it without caps I connect that cable to the new output pin header I've soldered to another point in PCB just before the output caps (see picture). The new IC is burning-in but, as I've got a cold and my hearing is well below par, I'm not going to evaluate the sonic changes at the moment. What I do recommend to you is to bypass C4, just for hearing the effect by yourself. In my amp and to my ears the improvement was not subtle.

And now a question for the "WNA-meister"

, i.e. Pink Floyd. Thank you for your previous answer, but I'd like more information about these diodes (I1, I2, D1 and D2), if I order them from other supplier than Rapid. I'm unfamiliar with diode terminology (I've seen references like 1N4148, 1N5711, 1N4007, etc., black or glass bodies, constant current, zener and other types, different values, ratings...), so any explanation and specific references would be welcome.

Best Regards,
Jose

 

[bludragon]

Hi Mike, thanks for your help. Haven't had a chance to play with it over the weekend. Will let you know if I managed to figure out the problem. At the moment I'm just running it with the rail splitter bypassed, but it does mean I have an ugly nest of wires hanging out the back hooked into a powered prototyping board.