Blueman2

It has taken me about a week to understand why so much of the music I am listening to on the KGSSHV Carbon seems distorted and overmodulated. Until last week, I had never heard of the term 'Dynamic Range' or 'loudness wars'. I was unaware of just how poorly much (most?) of the music recorded in the past 15-20 years was mastered. Not really a topic for this thread, but just thought it was worth mentioning as I wasted quite a bit of time and effort doing measurements on my amp to ensure it was correct, doing similar tests on my SRX-Plus, and then going back to my HD650s as a reference. Turns out it is the source material that was poorly mastered. In fact, this is a general issue with much of the music recorded in the past 15 years or so. One of the worst offenders is a favorite of mine: Beck. Morning Phase and Colors are both horribly over modulated and distorted. Luckily, there has been a re-master of my very favorite Sea Change which addressed the issue, but just a word of warning to those like me who are new to using Stax earspeakers with really good amps like the Carbon. If something sounds bad, it is probably the source material, not the DAC, Amp or earspeakers. Unfortunately, the source material is the one thing we have no way to address! Depressing.

Looking great! I see you are following mwl168's advice and elevating the resistors. Smart. Looking very good so far. So you will be finished my morning? I mean with the Scotch.

After more timing listening to both the SRX and the Carbon on my SR-507s, I am finding both amps reveal 'flaws' ins some of my favorite albums. Beck's Morning Phase is a good example. Parts of that album are way over modulated and distorted. Waking Light, at about 3:35 into the song, is a good example. That is the way Beck wanted it, so calling it a flaw might be the wrong word. But until I had these amps, I did not realize just how overmodulated it was. I will say that the SRX seems to handle the distortion better, giving a more pleasing representation. The Carbon just lays the recording issues bare and is not as pleasing to me. More accurate to what Beck wanted, but certainly not as pleasing. I have both amps on my work desk and each gets about equal use. Just depends on the mood I am in and what the source material is. Not sure where the hell I am going to fit the BH-BJT. New desk, I guess.

Yeah, I think I stayed up too late last night.Brain not working well today.....

James, have you (or anyone else) considered the option of increasing the output current loads above 7ma? Since I using an SE setup for my SRX, there is the option to greatly reduce the current for the o- (say to 2ma) and increasing the current for the o+. Given the 5 W limit for the 6SN7GTB, I think 12ma is possible. The combined would also be below 7.5W combined limit. Since I am not using the O-, might as well give more to what I am using, right? I got this idea after increasing the Carbon from 17 to 20ma and really noticing a difference at higher volume levels. Curious if this might provide good results for the SRX. Just thinking out loud really. Probably something I am missing in doing this. Would have to change some resistors, of course. Oh, and don't tell my wife I am even thinking this. She likes the SRX just as it is. Sssshhhh!

Well, the SRX-Plus is still getting love. My wife still likes its sound better than the Carbon. She says the Carbon is too harsh for her, and the SRX is quite petite compared to the humongous Carbon. I used every square inch of the SRX case, but the Carbon is only about 1/2 full. There are albums I like better on the SRX than on the Carbon. For example, my old favorite 'The Yes Album' just sounds better on the SRX than on the Carbon. Can't really explain why. Although my SR-507s are starting to get very worried that I am going to ditch her for an 009 Omega.

Of course, now I need to build a larger case for all these damn amplifiers. SRX-Plus, KGSSHV Carbon, BH BJT. Oh, and I definitely need a Stax Omega now as well. And a better DAC. Damn you all. If you give a moose a muffin.......

My wife already hates each and every one of you. It is about to get worse, though. Let me pan back on the Carbon build: And zooming in on the left: Guess my wife is not inviting any of you guys over for dinner anytime soon. Too bad. She really is a good cook.

Thanks congo5. I am going to order some very long Peek screws from the local dealer here. I plan to drill out the hole and put the screw all the way through the back and add a washer. Luckily, Solid Spot LLC, the main importer of high quality Peek screws, is just about 5 miles from me so I can just go pick them up. I changed my Carbon to 20ma (from 17ma) and at low levels, could not really tell much difference. However, when I cranked up the volume, it made a difference. Brighter and more shiny. In a good way, for my ears. I know it is cliche to say you can hear things you did not notice as much before, but in this case it is just true. I keep getting surprised by sounds I had not heard before coming from my old standby albums. I am dedicating all day tomorrow to just listening to the Carbon. I have no real desire to finish the chassis now; I just want to listen to it.

Oddly, Whitigir, they are only rated 300V. I got them not sure I would use them, but after visually comparing to the 400V rated ones I also bought, this one simply was better. More clearance from the board and better angle of the wires. Only time will tell if I have any problems, but I sure feel better with these after comparing to the regular 400V black ones from Digikey. And these are far, far away from the trace that a couple folks have had arcing from, so I cannot see that issue coming up for me. Part number 1546074-5-ND With my Carbon up and running, I just spent a half hour listening to Chris Jones' Roadhouses & Automobiles which is a really well mixed and recorded album. I have heard this album probably 100 times, but I was truly shocked by some of the sounds I was hearing. I kept looking around to see where the sound was coming from, but it was the recording. There were several times I was startled hearing things I just did not hear in the same way in the past. Definitely more sound stage, no question there. But also more punch and startling clarity of everything. I am running at 17ma right now, but will retune for 20ma tomorrow and listen some more. That is, if I can get to sleep. I want to stay up all night listening to music!!!!!

No, neither servo engaged. Just got lucky with initial setting. I played with the offset trim and was able to go up to around 25V and then easily down to 0 with a bit of drift. BTW, I did not populate the IC for the opto, just the OP27. Is the OP27 Servo 1 as labeled on the board? And is this the one that does not require the 10V to trigger, so I can just adjust to close to 0 and let it do its work? EDIT: I just finished dialing in balance and offset. Turns out the offset was not as good as I though initially. The balance was WAY off, at 75V. It appears that balance impacts offset, because after getting balance to 0, the offset was then off about 20V, but easily got that down to 0 also. On both boards, I am able to get balance to < 100mV with a bit of drift over time, and offset to < 500mV, again with minor drift over time. Both servos turned off (header not connected). I am now going to re-check the 17ma setting for the 4 outputs and let it run for about an hour. Then, depending on the thermal state, I will move up to 20ma and redial all the settings. One thing I will say is the amps put out a LOT of heat!! Really surprised by how much. Massive heat sink got warm all over after just 10-15 minutes of testing. EDIT2: I put the jumper on Servo1, and it works quite nicely!! The offset varies +/- 200mV around 0, but holds that range pretty well so far! Does the servo also help with balance? I did not think it did, but for some reason balance is also staying much more stable, going up and down within 200mV of 0. I think I am ready to listen to music. Finger crossed. EDIT3: We have music!!! And it sounds really, really, really great! First impression compared to the SRX-Plus is more clarity to the bass (sharper, punchier) and more clarity in the treble. I jumped a couple times hearing some high treble parts I had not really noticed before in a few songs. Everything just sounds sharper and more technical, if that makes any sense. I need to listen more, but there appears to be more sound stage as well. I am using SR-507s. I can't wipe this smile off my face.

Well, I decided since nothing was smoking or popping, I would give it 100% power. And sure enough, the offset quickly settled down to about 3V initially, and slowly went to about 0.4V after about 10 minutes. Great relief that everything appears to be working correctly, at least on the Left amp board. Now on to testing the other one and getting everything dialed in. Maybe tomorrow there will be music!!! EDIT: mwl168, you posted just as I was posting. I did not try that but should have. In any case, it is alive!!!!!! And if anyone noticed, I have a very strange looking header for B+/- rails on the amps. It is from Digikey. This item is nice because it allows the wires to be more up rather than across the board. But the real advantage is the way it lifts up off the board where the troublesome tracer is that a couple of people have gotten arcing from. See below how it looks. At least it allows me to see how much solder/flux there is under there.

OK, I will replace the washer right now. But I did do a quick test of the amp boards running at full voltage on the LV (using trafo plugged directly into the wall) and with about 45V (about 36% of full voltage) into the HV trafo using a variac. All 4 LEDs light up, but there was a strange reading I wanted to get advice on. The left DMM is showing the B+ rail. The right is showing the O- offset from ground. Both are about 150V. Does that seem right? Same reading on both amp boards. I wanted to make sure there was nothing obviously wrong before going to full power. Thoughts? mwl168, earlier you posted about your offset jumping to 170V on startup but quickly settling down after that. Could it be that the system needs full voltage to settle down? I have both servos off right now.

Thanks for taking the time to look the build over. I appreciate ANY input and feedback still being new at this. You are correct regarding the metal screw on the bias 10m90, but it does have a white plastic sleeve on it that came from an AVID kit. I also have the reddish brown rectangular insulators from the BOM in post #1. Is the rectangular insulator better? I would use PEEK here, but my PEEKs are M3 and the heat sink was not metric threads.

OK, if it ain't broke, don't fix it. I did spec everything to be able to do 450V. The trafo is really putting out 370VAC because it is spec'd with 115 primary and I typically run 125V in my house, and never less than 122V that I have ever measured.

I have the bias on both my SRX-Plus's at 570v, which drifts up to about 575-580 after an hour or so. I liked the bhultrimini psu which had a trimpot to adjust bias. Curious what the reason was for going with fixed zener string for bias on the GRHV vs the method used in bhultramini? In any case, I have a 120V zener in my cart for my next Mouser order. Going down about 10V (replacing the 130v with the 120v) should provide a good enough safety margin. Thanks for the guidance. With regards to the rail voltages (+407, -410), tell me if this is just being too OCD, but I do have a 10M resistor that I could parallel on top of one of the 390K resistors. By my calculations, that will bring that 390K down to 375K, so a total reduction of the 2x390 from 780K to 765K, a 2% reduction. So on the negative rail, it would come down from -410 to -402. Is that worth doing, or am I just being silly?

Actually, they said today they have one 25va 15V/15V left. Hopefully that ships tomorrow.

Pair of GRHV Fat PSU boards are up and running! It feels SO much less stressful bringing these up with a variac rather than just a plug&hope approach. All is fine, but voltages are a bit high. +407 and -410 loaded with a 15K 10W resistor for each board. Those test resistors get darn hot after about 5 minutes!! I am using a 350V trafo, so not worried about going above what it can feed, so might just leave that as is. However, the bias test point reads 590 and is still slowly climbing as the diodes heat up. Guessing I should replace one of the 150V diodes or the 130V with something about 10V less? Not sure what values are readily available so will have to look. I have also completed a pair of Carbon boards that @mypasswordis started but decided not to complete (he is going SMD route). I have pre-set the CCSs to 17ma. BTW, the process of using LV (in my case a pair of 9V batteries) to set the CCS was INVALUABLE. I could not set the current for one channel and found the cause: I missed a solder joint! Nice to find mistakes like that without having 800+ volts running through them! I have to wait until my 15/15 LV trafo arrives from Antek before firing up the amps (I am using the on board LV from the GRHV board until I decide if I want to built the GRLV). Those guys at Antek have cancelled my last 2 orders due to lack of stock. Never even let me know this time. I had to call and see why the order was late and they said they never shipped it. Ah, I guess that is why!

Thanks, Dr. Gilmore. So that is what the +/- points are on the board? What voltage is required to trigger 'on' state? In any case, I think I will leave unpopulated for now.

Another follow up on the timer delay. Is there any issue with populating the board with the 600R resistor and the CPC1117N and just not using it? Or do I need to jumper something to make HV come on in this case? I figured while I am building it, I might as well future proof the PSU.

Thanks once again, Michael. My next question was going to be the CPC1117N so you answered my question before I asked it!

I am about to embark on building a pair of KGSSHV (edit: I meant GRHV) Fat PSU boards (off-board heat sinks, as referenced in first post). I notice on the board silk screen it shows a 600R resistor, but none was listed in the BOM. I assume to 600R does need to be populated?

Question about filament voltages. I have seen numerous recommendations to slightly reduce filament voltage using an in-line 10W resistor of 0.5 Ohms or so. Looking at my Antek transformer, its filament windings are rated at 6.3VAC at 115VAC input. Here in SFO Bay Area, I average about 125VAC on my lines. This gives me about 6.8V on the filaments, so almost 10% high. I am not too concerned about lifespan of the tubes, so long as I get around 1000-2000 hours out of the output tubes (6SN7 in my case). Is it recommended I add a voltage reducer? For a pair of 6SN7s, would about 1 ohm be too much? I am not worried about the input tubes (12AT7s) as much (or should I be?).

+1 on the above. I started by buying an old SRD-7 off eBay and did the mod to give it 580V bias. Nice basic project and it provides a good sound, if not exceptional. Then, if you like that, move on to building the SRX Plus. That amp is pretty simple to build and is a big step up from the SRD-7. If I were doing this from scratch today, I would just go directly to the SRX Plus. Great sounds and really fun learning experience.

Wow, you are absolutely correct about clipping. But I thought it was more related to not enough power being provided by the USB PC driving the ODAC. But in any case, yes, I had to reduce volume (digitally) from the computer that is driving the ODAC via USB. I spent about 4 hours figuring that out. I actually ended up connecting a proper high end DAC built into my Integra receiver. It made everything sound much better. Now I know why!