[KimLaroux]

Thanks for the info.
 
I know PSU caps and signal caps have different requirement, but I'm still unsure what the difference is. I read that for power, you need low ESR and high ripple current. But what should I look for in a signal cap?
 
I just found http://www.partsconnexion.com which is a canadian company. They have all the bootique capacitors listed. And at nice prices too.
 
So I got Nichicon ES for cathode bypass and Elna RFS (Silmic II) for output caps. What would be recommended for power caps? (C1 and C6)
 
Edit: What about Nichicon FW for the power caps?

 

[Avro_Arrow]

Here is what I would use for C1 and C6
Panasonic FM
 
Digi-Key does not stock many values/voltages of the Nichicon FW 
so it's hard to compare. It should work OK, but I think the Panasonic
is a better choice.
 
Edit: Parts ConneXion didn't have many different values either...

 

[c12mech]

I've been using my ssmh for a while now and am thinking of doing something different with it.  I've seen a couple of active ground builds which I think would be cool.  What I really want though is a better power supply.  I ran across a thread in another forum about using a LM317 regulator for high voltages. 
 
Another thing I've seen is this regulator, TL783CKC, which can go up to 125v. 
 
Does anyone know of a good power supply circuit I could use for this. 
 
I know it goes against the whole "starving student" part of the name to do any of this but I'm not a starving student anymore. 

 

[bengoshi2000]

Well, I picked up a new set of cans (M-Audio Q40) which run at 64ohm.  Now I'm picking up some noise in my MSSH... primarily hiss.  Definitely independent of source.  Hiss does not get louder by turning up the volume knob.  At this point I'm guessing it's the RatShack 100k stereo pot that I used for the volume control.  On my next Mouser order I plan on getting an appropriate 50k pot and use the 50k resistors.
 
But let's assume for a minute that the pot is not the culprit... where else would I look to eliminate hiss?
 
I tried a different pair of 12AU7's... same hiss.
 
Hmmmmm.....

 

[HoopsMcCann]

Hello to all!  First post after lurking the past few months.
 
I finally finished building and debugging my 12AU7 starving student.  It sounds great and works as expected despite an issue I have with the MOSFET’s voltage divider (R2/R4 and R8/R10) that  has me stumped.   
 
I measure 48V between the top of R2 and ground which is the correct value.  However the voltages across R2 and R4 are only 27V and 15V respectively.  That totals 42V across both resistors so somewhere in between I’ve lost 6V.  The V measurements on both L and R channels are identical. 
 
What am I missing, where did the 6V go? 
 
I’ve removed the resistors from the circuit and measured them to be the correct values before putting them back in.  I’ve tried disconnecting the voltage divider from C2 and R3 to isolate it and get the same results. I even changed my DMMs battery in case it was low.
 
Nevertheless, with the voltage at the Gate only 15V instead of the designed 17V the amp still works. 
 
I measure 13.3V across the heaters but that is probably unrelated to the issue.
 
At this point I am inclined to leave it as is since the amp has been running without problems and I am already enjoying the music.

 

[KimLaroux]

Hi guys. I'm still salvaging parts for my MSSH. I got two options for the case and I thought i'd ask you guys what would be best suited.
 
The first option is to use one of those project boxes I've had laying around for a few years. They are about 1.5 X 4.5 X 6 inches. It's a "tube" design, with a main body made of 2mm thick aluminum. It opens on both ends with plastic covers.

 
The other option is to use the body of an old multimeter. This multimeter was given to me this week end, and since the dial was broken, it couldn't be repaired. The case is about 2 X 5.25 X 7 inches. It's made of thick plastic, with metal screw mounts. The case's faceplate opens up.

 
So here's what I think. The aluminum case is smaller and might look better once finished. It has a nice mate finish except for the end covers which are shiny. All the walls are square to each others. The problem is it opens only on both ends, making it a pain for P2P if I want to mount the MOSFET and tubes on the top of the case. I'd have to cut the bottom off and then figure tidy a way to mount it back on. The heat sink is also to wide and I'd have to cut the two end walls off... but then as those are used for mounting, cutting them off will just make the eat sink look better.
 
The multimeter case is heavy duty and made of thick 1/4 inch plastic. It will obviously be able to withstand abuse, but abuse and lamps just don't really go together anyway, so this point is not really relevant. Since the case opens up top-off, it makes P2P easier. It's also larger, with more inside room and place on top to fit the whole heat sink. It's shiny black finish with white engraved lettering. I believe this option fits well in the Starving Student philosophy. 

 The problem is the external walls are angled, which would make stuff mounted to them look weird. Internally the walls are square though.
 
Ah decisions... I hate making them. So what do you think? Is there anything I am unaware of that would make one of those case better suited for this amplifier?

 

[RanDiesel]

Hi, I have been working on a SSMH for a couple weeks now and have finished the case work and starting working on the circuit. But I have never built a circuit like this before and in fact this is the my first time even soldering. I was having a hard time laying out the resistors and capacitors for a point to point circuit so I found these little break apart pcb boards at radio shack and laid out the right side of the circuit and soldered it together. I was able to fit all the components of the right half of the circuit on this little board with the terminal strips, so I should just be able to screw in the wires from the sockets and mosfets. But I am wondering if it is bad to have the power and signal paths so close together? I was also wondering if it is ok to run all the grounds together and then just ground the single wire? So before I build the other half of the circuit I was wondering if anyone could comment on this layout, I don't really want to build the whole amp only to have a hum or distortion because I made a rookie mistake in the layout. Here are the pics, any help is appreciated.


 

 

[bengoshi2000]

On my MSSH I grounded to one main ground wire with no I'll effects. I didn't worry about where I ran the sound/power wires. The only noise I've ended up with came from the cheap azz ratshack stereo pot I used.

 

[KimLaroux]

It is best to connect all your grounds to one single point. This is called a Star ground. The idea is to avoid having a chained ground line where one wire goes around the circuit connecting all the grounds. The problem with this point to point ground is that the current going trough the ground wire an passing by all the junctions will create hum and other artifacts, like not having a 0v ground. Think of the ground wire acting as a voltage divider. Since it's not 0 ohms, there will be a voltage drop inside the wire. So at each points along the ground wire you have different voltage. This voltage will vary depending on the current passing trough it. Dirty.
 
Search images for "star ground" if you are of the visual type.
 
But then, all this might be over kill for this design. It's not like you were building a 50w discreet class A monoblock.

 

[RanDiesel]

Thanks for the info, I am going to go ahead and finish the other channel and put it all together. I will post pics when it is done.
 
I have seen several posts about putting a delay in the circuit so that you don't have to unplug your headphones before turning it off and on, but I haven't read about anyone implementing one. Anyone have any success with this?
 

 

[switch32763]

3 years ago, I bought all the parts for a 19J6 Starving Student build. I had a working breadboard prototype, but for some reason I just never saw it through to the end.
 
Fast forward to now, and I've been working on it again. Unfortunately, the other day I was trying to isolate a hum when I accidentally shorted pins 3 and 4 of the tube with my multimeter. The tube immediately got a lot brighter than the other channel's tube, so I shut off the whole thing. After some testing I realized it was actually the MOSFET that was shot, not the tube. So now I'm back in business as long as I get a new MOSFET, which is obviously really cheap. But I'm having some doubts about using my rare 19J6's at this point, since I already thought I killed a tube. I'm thinking about switching over to a 12A_7 build since this would be the perfect time before I actually wire everything and drill the enclosure I plan to put this in.
 
So basically, I have a few questions:
 

• Is it worth it to stick with the 19J6s for any reason? The only reason I'm considering switching over to 12A_7s is for when I lose a tube in the future and I have to purchase a new pair. I figured with 19J6s I'd be out of luck, and most likely have to also re-drill a new enclosure to fit the 9 pin socket for 12AU7s.
• Assuming everything goes well, for how many hours of use can NOS 19J6s be expected to last?
• I know for sure that I fried a MOSFET, but is there a chance I did damage to anything else? Specifically the tube, because it doesn't look there's anything else that could be shorted to ground by shorting pins 3 and 4.
• From what I could find online, it seems like 9 pin sockets typically fit through a 7/8" hole for the socket. Can anyone else verify the size of the actual socket part of their socket? The diameter of my (7 pin) socket is 0.610". Also, what's the typical center-center distance of the mounting holes for a 9 pin socket? For my 7 pin, it's only about 0.89".

 
Thanks so much to anyone taking the time to read this!

 

[KimLaroux]

I just placed an order at Parts Connexion. They had a 20% off special on every electrolytic capacitors, so I went all out and got the best they had to offer.
 
I finally got my hands on some TL783 by Texas Instruments. The datasheet for this part is very verbose, so I feel comfortable about using it to build a linear regulated PSU for the SSMH. I studied many headphone amp's PSUs and noticed that they are more or less all built the same way. I'll just build mine according to that. I was lucky and found a transformer with the perfect voltage for this chip. Salvaged it from a vintage stereo. I'll let you guys know how my experience goes.
 
For those who prefer to stay within sight of the beaten path while upgrading their SSMH's PSU, take a look at what I found:
 
TL783 based adjustable 700mA PSU. ~25$ free shipping

 
It's a nice little unit sold on Ebay. Searching for "TL783" brings up a few listings. The item's description even includes a schematic and some PCB traces. You'd need a transformer with ~60v of secondary voltage. Can go higher, but not really lower. I'd like to listen to this thing paired with the SSMH. At 25$ + a transformer, I think it's expensive for something like the SSMH. It might be worth it though if you can salvage a used transformer... or a switching PSU with at least 60V.
 
Audio-gd sells discreet PSUs. Obviously these are much higher quality than the unit sold on Ebay, but if this translate directly to better SQ... I don't know. Here are two that could power the SSMH:
 
PSU-L 500mA 12$ + shipping

I'm not sure if this one actually goes to 48v, but I guess it does. I sent an email to Audio-gd asking. Wow! Kingwa answered while I was typing this! This unit can supply up to 60v. This one doesn't have a diode bridge to rectify AC, so you'd have to add this before the input. You'd need at least 50v at the input. Note that you could use this unit with a 50v or more switching PSU. At this price, I think it's a good value to use with the SSMH. And since it has only 2v of dropout, it can be used with a much lower input voltage than the TL783.
 
PSU09 35W 32$ + shipping
 

This is actually a dual rail power supply. It might be overkill for the SSMH, but it's still an option. You'd actually only need half of it for the SSMH.

 

[KimLaroux]

I have a feeling this thread is dead... has the SSMH fell out of favor for cheap and easy DIY project? 

 
But anyways, I've spent the lass week working on my amplifier, and I thought I'd take a break to share my progress with whoever follows this thread.
 
I ended up building a custom enclosure from pieces of clear plastics I had in stock. The top and (yet to come) bottom are from a 5/16" thick clear plastic I salvaged from an LCD monitor. I think it's PMMA (aka plexiglass), but I'm not certain. The sides are 1/8" polycarbonate (aka Lexan). The whole thing was cut and filled by hand. The only power tool I used was a hand drill. The top, front and back panels were frosted using sandpaper. This adds a sens of focus to the build. A completely transparent case is just too transparent, it makes the build look like a mess.
 

 
It took me a few hours just to mount all the hardware on to the case. After breaking two screws inside the plastic, which I had to drill out without damaging the original hole, I had the brilliant idea of cutting a V into one of the screw and use it to make fillets inside each holes. Yes, I had to break two screws before I finally gave up. All the metal parts are grounded to earth. The electronics are isolated from the heat sinks.
 

 
Here's a picture of the progress I have made so far. The AC side of the power supply is done. There's a mess of wires in there. I didn't want to cut the wires from the transformer too short, just in case I want to re-use this transformer in another project. It's a 100W transformer with two isolated secondary windings, one of which has 5 poles. The two blue wires are a 6v secondary. I'll use this one to power the LEDs in the tube sockets. 
 

 
I'll be working from there. I should get the 48v rail up and running tonight. Any input would be greatly appreciated. If you spot a problem with my design, something that could potentially go wrong, I'd really like to know. Thanks. 

 

[KimLaroux]

I finished the power supply and started distributing the power across the amplifier. I'm now stuck at a question mark regarding R13.
 
I use 4 decoupling capacitors instead of 2. I solder them directly to the item it decouples. I also try to separate the channels as much as possible to reduce crosstalk. At this point, dividing R13 into two resistor would make more sens in my design, since I'd just solder a different one from the central power rail (like, physical rail...) to each side of the amplifier. The other option seems dirtier, since I'd have to run a wire from one side back to the next....
 

 
So I'm wondering, would it be better to use two different resistors instead of just one R13? Is it irrelevant? Or is it a bad idea?
 
I searched for information regarding R13 and it's role in the design, but couldn't find anything, so I'm asking. My guess is it's used as a CRC filter to isolate the MOSFET power from the rest of the amplifier. Is that correct? If so, then wouldn't using two, one for each channel, also isolate the channels from each others?
 
Obviously I'd have to find two identical resistors. But that's not a problem, unless a difference of less than a single ohm have adverse effects. 

 

[proid]

Quote:

I finished the power supply and started distributing the power across the amplifier. I'm now stuck at a question mark regarding R13.
 
I use 4 decoupling capacitors instead of 2. I solder them directly to the item it decouples. I also try to separate the channels as much as possible to reduce crosstalk. At this point, dividing R13 into two resistor would make more sens in my design, since I'd just solder a different one from the central power rail (like, physical rail...) to each side of the amplifier. The other option seems dirtier, since I'd have to run a wire from one side back to the next....
 

 
So I'm wondering, would it be better to use two different resistors instead of just one R13? Is it irrelevant? Or is it a bad idea?
 
I searched for information regarding R13 and it's role in the design, but couldn't find anything, so I'm asking. My guess is it's used as a CRC filter to isolate the MOSFET power from the rest of the amplifier. Is that correct? If so, then wouldn't using two, one for each channel, also isolate the channels from each others?
 
Obviously I'd have to find two identical resistors. But that's not a problem, unless a difference of less than a single ohm have adverse effects. 

 
What a great build 

. I really enjoy my MHSS and i'm planning to do another one ptp with better part and regulated power supply, your posts are really interested. So which psu board did you use?