|
|
Post by krojules on Sept 7, 2012 8:51:04 GMT
Yes I already set the trimpot who has an effect on the 1k5 resistors (1.35v).
You speak about this work or about another thing ?
Do you think I have to post a picture of my amplifier board to see something weird ?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 7, 2012 9:04:16 GMT
Yes I already set the trimpot who has an effect on the 1k5 resistors (1.35v). You speak about this work or about another thing ? Do you think I have to post a picture of my amplifier board to see something weird ? The trimpot should be set to give close to the same resistance from the centre to both sides, and then those voltage readings across the nominated resistors written down and posted here for analysis. Frans is experienced with the Panda, so any variation there may help him /us decide if the 2nd stage is O.K. and perhaps indicate if the input FETs are reasonably matched or not. Alex |
|
|
|
Post by krojules on Sept 7, 2012 9:48:29 GMT
Here is measurement (if I understood what you want).
You have to read these informations with the Alps at the top of the board. Lettering on the board as you read a book (I told you my english is poor).
Problem side (the right part)
K246: -16.3v/-16.96v (center pin)/ 0.958v)
2SK246 UP: 0.958v/0.00 mV (center pin)/ 15.55v
DOWN: 15.55v/ -25.2mV (center pin)/1.16v
1K5 resistor UP 1.358v
1K5 resistor DOWN: 1.355v
Good side (the left part)
K246: -16.61v/-17.13v (center pin)/ 0.951v)
2SK246 UP: 1.093v/0.00 mV (center pin)/ 15.78v
DOWN: 15.78v/ 0.6mV (center pin)/1.014v
1K5 resistor UP 1.355v
1K5 resistor DOWN: 1.355v
voilà.
I hope these informations will help you a little bit more.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 7, 2012 10:08:35 GMT
On the measurements so far, it suggests to me the possibility of poorly matched Input FET pairs.Is the trimpot set for equal resistance both sides ?
I will leave further analysis to Frans, as he has experience with the Panda.
Alex
|
|
|
|
Post by krojules on Sept 7, 2012 11:52:12 GMT
I set the voltage to the trimpot at 1.5V and now i can reach minus 17.6mV !
If I set to 1.2V it is worst (better was minus 200mV).
When I turn the trimpot, I heard the relay switching on and off.
Someone could explain me the process ?
If I keep this voltage, Does it damage the amplifier with the time ?
It is really interesting to control all of these component to understand electronic.
|
|
|
|
Post by parrymike on Sept 7, 2012 13:35:33 GMT
What type and power rating do I need for the output resistors? I've had a look on ebay for some decent 2w composites (~100 ohm) but I had no idea they would be soo exensive! Are 1/2W metal films ok to use? I suppose I could parallel them to increase power rating?
Mike
|
|
joethearachnid
Been here a while! 
Old head on young shoulders.
Posts: 380
|
Post by joethearachnid on Sept 7, 2012 18:45:55 GMT
Krojules:
I think most of us ended up using 1.5V using the trimpots. It was agreed that it sounded better and it shouldn't damage the circuit since it's well within the circuit limits. I'm sure Frans would have pointed out if the voltage had been high enough to cause any damage.
parrymike:
I don't actually know what the best rating would be - I know that the Panda can output quite a bit of power so I used three 1W 330ohm resistors paralleled (and carefully matched for the same total) for 110ohms as output resistors. That might be excessive, but it's worked without problems.
-JoetheArachnid
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 7, 2012 19:55:18 GMT
It was me that suggested the 1.5V in the first case after fixing someones Panda b.t.w. 1.5 to 1.6V is about optimal and the 1.2V is simply too low and the amp is NOT in it's operating point at that voltage. I have no idea why the desiner specifies 1.2V ...  The procedure is to set the 1.5V on the 1k5 resistors and after this to set the DC offset voltage. You can measure this at the points mentioned earlier in the thread. Once the relay is energized you can also measure DC at the output connectors. IF the input FET's are well matched the offset potmeter will be in the middle setting (can only check this by turning it all the way to one side till it 'clicks' and count the rotations in the other direction (pots are 20 turns so 10 rotations would be the middle) When the FET's aren't well matched (this should have been done before the assembly as instructed by HiFiDIY themselves) the DC offset potmeter will also be 'off centre'. Perfectly matched FET's out of a relatively small sample size (only a few extra FET's were included) will be more a question of luck. Let's just say the DC offset potmeter isn't a multiturn pot for nothing. The measurements seem to confirm the PS is O.K. (although you seem to have mixed up 2 values in Q3: 16.9/20.6 and 17.56 for center pin and think you meant to write: Q3: 16.9/17.56 and 20.6 for center pin, but you are forgiven). The second set of measurements seem to confirm the obvious (as Alex also mentioned) the input FET's on one channnel don't seem to be matched close enough. When you are happy with the sound and you can get the DC offset down to -15mV it'll be fine. It is much better than -150mV which you do not seem to have noticed. Personally I would have bought some more 2SK246 FET's and look for some closer matching pairs and install these. About the output resistors: Personally I wouldn't be anal about it and would gladly use wirewound 1W or 2W resistors as resistors are linear elements and they do not introduce distortion or colouration (they are physically not able). Bandwidth limiting is not going to happen with the impedances involved and would stay well above 100kHz. Yes... I know I am wrong and many audiophiles can clearly hear the differences proving my lack of knowledge and ears on the subjective side of audio... but that's simply how it actually works. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 7, 2012 20:52:44 GMT
Krojules
Just to elaborate on this point from Frans. Although the schematic has the input pair devices labelled differently, they are both 2SK246 FETs.
Alex
|
|
|
|
Post by krojules on Sept 8, 2012 7:36:28 GMT
Hello,
thank you for answers and advices.
If there was no trouble when I was listening with my headphone, it is because the amplifier were adjust to 1,5 volt at the first time. I decided a few day to adjust DC offset and decrease voltage drop across 1k5 resistors. By decreasing the voltage drop, I moved the DC offset in the same time. Maybe the amplifier were in a good setting before !
I'll try to set to 1,6 volt and obtain a zero (close to) offset.
I heard your advices about unmatched 2SK246 and I'll purchase 5 pairs at desco know oent price.
I've got tools to unsolder/solder component at work.
For the trimpot, I'll try to count number of turn before the relay click to know where am I.
Thank you again for your help and advices, I really apreciate.
Next episode when I receive FETs.
Julien
|
|
|
|
Post by krojules on Sept 10, 2012 7:13:12 GMT
The News.
This week-end, I increased voltage to 1.6 volt.
Then after, I was able to adjust the DC offset close to zero mV on the both side !
I really don't know if I have to buy new Fets now ?
Will it theoretically improve the sound ?
I'm really please to listening music with this amp.
For me, it is a good combination with my AKG K271 MKII
Krojules.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 10, 2012 7:29:59 GMT
IMO, I doubt that it will do anything more than help prevent long term deterioration to your headphones due to the original -150mV DC offset, unless the balance trimpot is now close to centre adjustment. i.e. equal resistance on both sides of the centre pin.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 10, 2012 19:04:30 GMT
I really don't know if I have to buy new Fets now ? Will it theoretically improve the sound ? If it works as expected you don't have to buy new FET's. IF it were my amp I would install some matched FET's as theoretically (the FET's seem to be spread quite far apart in characteristics) you could lower the distortion a tiny bit. You can't go below a certain value of distortion determined by the amp design so the differences will be marginal. The headphone you have will have at least 100x higher distortion than the amp has, even with unmatched input devices. The K271 distorts between 0.3% and 1% which is pretty good for a headphone. The Panda will easily reach 0.01% |
|
|
|
Post by pandapops on Sept 20, 2012 23:09:12 GMT
Are you are that the trimpots are good? I just did a new Panda and had a bad 200 trimpot so the DC offset could not be adjusted properly. The pots have 3 legs, two of them had continuity instead of resistance.
|
|
|
|
Post by PinkFloyd on Sept 21, 2012 10:56:21 GMT
I've got a couple of these kits left if anybody is interested.... £50 shipped.
|
|
mrarroyo
Been here a while!
Our man in Miami!
Posts: 1,003
|
Post by mrarroyo on Sept 22, 2012 12:42:54 GMT
Mike, any chance you might be coaxed into building one of the kits including your magic touch of pixie dust? If yes please send me a PM, thanks.
|
|
|
|
Post by PinkFloyd on Sept 22, 2012 14:15:06 GMT
Mike, any chance you might be coaxed into building one of the kits including your magic touch of pixie dust? If yes please send me a PM, thanks. Hi Miguel, I'm more than happy to build the kit up for you but it would be more of a populated "ready to go" board than a fully built up amp..... I'd love to build you a RTG amp but the shipping to the USA would be LOTS.... let me weigh the amp I have just for an idea..... I'd populate the board for you for free mate, would just be a case (excuse the pun) of somebody casing it all up your end. I will be selling Panda Three (you can see Panda three here: rockgrotto.proboards.com/index.cgi?board=m&action=display&thread=6604 ) it's come on leaps and bounds since then, with matched transistors and what have you, I'll do you it at a very good price Miguel but we are looking at £70 GBP just to ship the bastard.... anything over 2 kilograms (4lbs to you and I) means RAPE when it comes to shipping overseas by Royal Mail..... actually, it equates to daylight robbery! Little Pinkie was designed to be "shippable" across the globe at a reasonable cost.... BIG Panda (fully built) is not so user / wallet friendly when it comes to shipping. |
|
mrarroyo
Been here a while!
Our man in Miami!
Posts: 1,003
|
Post by mrarroyo on Sept 23, 2012 12:05:27 GMT
Thanks mate, I look forward to your PM. Cheers!
|
|
|
|
Post by st on Oct 9, 2012 19:30:07 GMT
Hi, how much for Panda 3, and can you supply it with a case, I am in SW Scotland
|
|
|
|
Post by st on Oct 14, 2012 17:34:31 GMT
Hi did you see my last post?
|
|
|
|
Post by st on Oct 27, 2012 16:23:59 GMT
anybody in here !!!
|
|
mrarroyo
Been here a while!
Our man in Miami!
Posts: 1,003
|
Post by mrarroyo on Oct 28, 2012 12:04:49 GMT
You may be better served by sending Mike (Pinkfloyd) an email or PM.
|
|
oldson
Been here a while!
Older than a well used glove
Posts: 414
|
Post by oldson on Apr 2, 2013 18:31:08 GMT
 i have recently bought the superb panda (in the pic) from Mike, and am so impressed by it that i am considering selling my X series gear to reduce clutter. at the moment my x-psu,x10-d and x-can are used only for my sacd player, which i only use on rare occasions. my question is-- how easy would it be to mod the panda to have 2 switched inputs? in the meantime i do have a switchbox (from maplin), but this adds to clutter again and defeats the object somewhat. cheers |
|
|
|
Post by PinkFloyd on Apr 2, 2013 20:07:15 GMT
Fit a switch, connect one side to the input and the other side to the input........ pretty easy  IN / Out / IN Mike. |
|
oldson
Been here a while!
Older than a well used glove
Posts: 414
|
Post by oldson on Apr 3, 2013 12:38:53 GMT
Mike
just got some phono sockets from maplin and they look shite compared to the ones you fitted.
could you please post a link to a supplier?
cheers
|
|
IN / Out / IN