Modifying the Xiang Sheng 708B [56K!!!!]

[sonq]

I was thinking if all this modding is worth the time and cost of components - may add up to be almost same or even more than the cost you've saved by not buying from Woo or RSA.

Looking at the incredible workmanship in your photo, you must find much pleasure & satisfaction in modding.

[nikongod]

Originally Posted by sonq

I was thinking if all this modding is worth the time and cost of components - may add up to be almost same or even more than the cost you've saved by not buying from Woo or RSA.

Make a parts list, and see what the upgrade parts cost is to build an amp like this using parts of the same "grade" as is found in a Woo, singlepower, mccalister, decaware, or RSA amp.

According to the "mot" rules, DIY'ers are not allowed to assign a price to hourly labor, and must sell their work purely at parts cost. If you would like to ignore that, the labor rate is $25/hour based on what I guess a solder tech who can probably solder better than you gets paid. If it takes you 2 hours to unstuff and repopulate this board you have it bad or are likely P2P'ing the amp which is not required if you use parts of similar quality as mentioned above, and should only account for $50.

Now then we have the finished product, with a base price of $120 which is about the average used price for the amp , the upgrade parts costing about $75, and $50 of labor (which we should not even count, but I am initiating a premptive strike.) This is just a hair under $250!

As far as SQ is concerned, this is a very subjective matter. Try it against something from RSA at the $250 price point used LOL. Hornet M FTW! Im not sure if any of the woo amps ever sell in the $250 range, I dont think they do.

[glitch39]

The 708B is a true sleeper. Earlier in this thread, it was noted that this is almost a clone of the MJ amp, which is very highly regarded

Modded, the SQ rivals the 336 and beats the figaro, both within striking price range

Not the same league and performance level as my woo 3+, however

[Stephen_Ri]

For several months now I've had to apply leftward pressure to the the end of my headphone cord while it is plugged into the 708B in order to hear the left channel. The problem is not with my cord. Recently, the left channel has stopped working entirely. Any suggestions? How difficult is replacing the jack input, and what reasonable options are avail?

[Random Murderer]

Originally Posted by Stephen_Ri

For several months now I've had to apply leftward pressure to the the end of my headphone cord while it is plugged into the 708B in order to hear the left channel. The problem is not with my cord. Recently, the left channel has stopped working entirely. Any suggestions? How difficult is replacing the jack input, and what reasonable options are avail?

it might not be the jack itself, but the front aluminium plate. it seems to 'drift' forward a bit and cause the plug to not be able to be fully inserted into the jack. try fixing that first.
the jack itself is not hard to replace, the real challenge is either finding a jack with the same pinout or an open jack that's long enough to fit through the aluminium panel.

[Random Murderer]

i just fell in love with my 708b all over again. i replaced the crappy wima film caps today(mine weren't the mentioned 400v 220nF but 630v 220nF, they must've run out of 400v) with some cheapy Evox 400v 220nF($.25 a piece) caps i got at a local electronics store, and even THESE are a ton better than the wimas. i highly recommend that everyone do this if you haven't already, like i said, even my cheapo caps are better than the wimas they used. the evox caps sound a bit more airy and natural, and i actually found that they calmed down the outrageous bass a bit and extended the high end a bit more.

and just to reiterate:
SPEND A DOLLAR AND REPLACE THE WIMA CAPS!!!

[nikongod]

This evenings festivities involved replacing the plate "resistors of sound quality" on the input/gain tubes with constant current sources.

For quite some time I have been using the stock 33k ohm resistors here with a LED cathode bias and everything was pretty good, but why not mess around with perfection? In my amp, the 33k resistors put about 105-110V on the plate of the input tube which agrees with the voltage in the headwize MJ article. As I have been researching the 6DJ8 I have come to realize that it is a fairly dinky bit of metal and decided that running the output tubes at rated power (per section!) may not be the best idea anyone has had in a while.

I used 10m45 CCS made a few calculations for the resistor value and got VERY close to my desired result.

I took this opportunity to reduce the current in the gain stage by about 10% (from 2.2ma to 2ma) which reduced the plate voltage on the gain tube a few volts, and since the plate voltage is "echoed" through the output stage the current in the output stage dropped a little too! My choice to reduce current in the gain stage comes with the fear of less linearity, but my thought process says that the CCS loaded tube at a slightly less linear operating point is MORE linear than the resistor loaded gain tube.

[Cloth Ears]

Finally, an update to my little XS headphone amp (previously mentioned in despatches here in this post).
Looks the same on the outside, right?

But there's been a few changes on the inside. PS board has been mounted edgewise (and runs cooler as a result). Bleeder resistors added to the PS caps. New wiring. New headphone jack. New RCA's. ALPS volume control ($8!!). 3 new output tubes (PS are still original). Blue output caps are probably Philips. Other caps are EROMAK (gold), plain-jane Mundorf (white - very neutral, we like them), Russian PIO (green), and I'm blowed if I can remember what the black cap with the red ends is (maybe a Wima?).

And here's an overhead shot:

And a close-up of the changes at the front:

And the crossfeed circuit (which we have to mount outboard, as it just doesn't seem to work properly when we tried to mount it in the chassis):


James did all the work, I just scrounged around for some of the parts. He may be able to answer any of the complex questions this might arise...

Sound, well it's still pretty fat at the bottom end, but much better controlled and very lacking in distortion. Top end is much better then previous. The mids were always quite good and we have only minor gains there. Volume control is about 12 o'clock for 'quite loud now, as against 8 o'clock previously.

About $100 spent on parts, of which about $50 made it into the chassis. James spent countless hours playing around with the set-up and had numerous insights into how he's going to build his own headamp.

[Random Murderer]

I'm thinking of adding in another 6DJ8 as a buffer, using one triode for each channel.
Schematic as follows:





Ideas? Comments? Suggestions?