Discussion in 'Exclusively Sansui' started by HDJR808, May 10, 2016.
Yeah that's why i quoted your post. Isn't that funny? at least?
Funny, yes. Looking at the operation of the meters, this is my amateur's guess about how this was set up:
I am assuming that the raw meters are voltage meters, and that part of the purpose of the meter driver boards is to reduce the voltage level to a point to not overdrive the fine mechanisms in the meter itself. But also, looking at the 2, 4, and 8 ohm voltages of the transformer windings, they are not a linear variation into those loads, and I am guessing that the ic circuitry on the board is a means of making the meter perform in a linear fashion for the three maximum output voltages. And while the Odb and -40db levels are adjusted with the 49V and 490MV references from the outputs of the 8ohm tap, the 4ohm tap is used to drive the circuitry and the range is interpolated for the 8 and 2 ohm output voltages.
I could be completely out to lunch on this, but as I said, that's my amateur's guess.
Hopefully tomorrow I will get some more of the wiring harness completed...
Today was the replacement of the line level coax cables from the buffer board to the driver boards. On the originals, the terminal housings had been removed, and then the terminals were clipped and soldered directly to the header posts. This is one of the two new leads that I made up:
The terminal housings and terminals are new, and the coax is some high quality 26awg cable stripped from a multi-conductor RGB video cable. This coax has a 16.7pf per foot capacitance, which is good spec for line level audio interconnects, but to be certain there were no anomalies I made up some interconnects with it that I have been running between my amp and preamp for a few months now and they sound fine.
The original harness uses a single terminal housing at the buffer board that is polarized, and the two leads split from there:
I had no luck finding that particular terminal housing, but was able to use the same housings on the source end of the cable as are on the driver boards and just install them side by side:
Now it's time to figure out what to do with the power switch. The original appears unmolested on the outside, but considering the condition of everything inside this amp my guess is that it's pretty sludged up internally. I'm just not sure that I want to try and crack it open...
As long as its not that old Kapton wire that shorted out and caused plane crashes due to brittle and chaffed insulation.
I had never heard about (and kind of wish I didn't know now ) the Kapton aircraft wiring issues....
But no, the wire I have used is Tefzel, in the Teflon family, that appears to be one of the wires of choice to supplant the Kapton wiring that has been implicated in aircraft failures. The Tefzel is supposedly a tougher sheath than Teflon, and the 22759 spec is tinned as opposed to a lot of the silver plated Teflon wiring.
Before I dealt with the power switch, I wanted to go ahead and install the fan thermostat and wiring. Using the 240V fan, I could eliminate the 120ohm series resistor, and also, since I did not intend to enable the auxiliary AC outlet, I did not need to tap into the fan circuit for that. This made the wiring a little cleaner:
The leads for the fan were sheathed in some silicone tubing and terminated with the appropriate .125 Faston connectors:
The beefy 20 amp power switch was physically in OK shape, but testing for continuity, the contact was intermittent with any jiggling of the lever. Unfortunately, not only does the bracket for the lever mechanism make it difficult to disassemble the switch, the plastic housing has some type of sealant applied where the terminals exit and I didn’t want to disturb that.
Like many other Sansui power switches, this one is a double pole switch that is wired through only one pole, so I have the option of utilizing the untapped pole. Since I didn’t want to have the lever operate in reverse, I figured I would try and see if flipping the lever mounting in the bracket itself would actuate the other pole, but in the proper “flip up to turn on” activation.
Using my trusty old ignition points file, I carefully filed down the swaged end of the lever pivot pin so that it could be removed. With a few scratches on the bracket, but not too much the worse for wear, the lever was flipped, and the alternate pole operates as normally open with the lever in the down position.
With a new X2 safety capacitor, the switch was wired up with (again, thanks to Apex Electronics) some nice hi-temp SFF-2, SEWF-2 appliance wire. This 14awg BICC brand tinned copper wire is UL approved, rated for 150 ℃, and has silicone rubber and fiberglass insulation. It is very flexible and easy to work with even in the 14awg.
These are the jumpers that I made up for the voltage selector:
And with jumpers installed:
Very well done. Congrats.
Amazing and commendable attention to detail - really excellent.
Really nice work! And thanks for sharing it with us.
Thanks, I appreciate everyone's positive comments.
Things are getting close to completion I think. Finally got a chance to get the driver boards installed, and unfortunately encountered a pretty big vapor lock on my part. In the category of "what was I thinking?" I realized that when I found what I thought would work for new terminal housings for the 5 position header on the driver board, I didn't take into account the pin size on the headers. I was so concentrating on getting the spacing correct (a complete pain since that style is obsolete) that I completely neglected the fact that the terminals needed to accommodate the 1.14mm diameter pins. Turns out that they didn't.
So, the solution was to trim two separate housings to fit side by side and mate to the polarized spacing of the header. Luckily I had housings that provided correct spacing, but it meant that I had to remove the heatsink and re-terminate the leads coming off of the connector board and TO-3 sockets. This sure underscores the old adage of measuring twice and cutting once, but at least the boards are in and the connections seem to be sound.
One thing that I have found is that working with most of the replacement wiring has been a little fiddly during reassembly since I have tended to cut everything a little bit long just to make sure that nothing comes up short. Things look a little untidy, but at least this has prevented any of the harness coming up short.
At this point, it looks like all that remains is installing a new power cord and all of the panels and faceplate. I have some ideas for putting together a harness for the meter leads for setting the bias that will allow me to run two meters and check both sides at once, bypass the meters for the inrush surge and, and keep the top panel closed while readings stabilize. I also got a 250w bulb for the DBT that I'll have in line with the variac for initial startup. Hopefully that is adequate wattage.
Things are looking rather complete now:
Man you never seize to excite and surprise me.
As for the harness you'll need some kind of switch somewhere so to switch modes , between normal operation and bias settings. Also you'll need to fix climax scale , as to what readings on the meter represent which bias current.
Yes, this is the jig that I completed today:
The cheesy little knife switch (somewhat mangled in shipping) allows for the meters to be bypassed upon startup, protecting the internal 500ma fuses, and when opened directs the current to only the meters. The barrel connectors are tied to the internal measuring points, will be removed once the bias is finalized, and allow the meter harness to not get tangled up while manhandling this behemoth of an amplifier.
No pictures yet, but there was a little more reassembly today, with some of the frame being bolted on. Hopefully, all will work properly and it will sound good, but, even cleaned up, it seems destined to truly be an ugly duckling on the exterior. With all of the dents and scratches and rust and corrosion, it will always display the marks of its abused history.
...Thank You for the very well done Report , Sir - Keep my fingers crossed for the fire-up of this remarkable reBuild from scratch...
Well, unfortunately, my fingers are still crossed, but no joy so far...
Everything appeared to be complete and in order, so a couple of days back I figured it was time for the flame test. Using a variac and a dim bulb tester, I tried bringing it up to voltage, but the 250w bulb refused to dim. Poking around, I found that the 16awg power cord that I had installed was apparently shorted. Don't know where that cord came from, but the hot leg read as open and I'm guessing it was internally shorted since the bulb was at full brightness. Since I tend to never throw anything away, this was just some random cord out of a bin of stuff and appeared to be relatively new. Anyway, it was odd that it was faulty, but dummy me for not checking it first.
So, new cord, DBT dims, and I set the bias. So far, so good. Things seem stable, so I hook up some speakers and a preamp and get nothing. Start checking voltages to the boards, and all look good until I get to the buffer board, and find that I had completely overlooked the power leads from the transformer. Not ashamed to admit that this was a complete dumbass vapor lock on my part. Guess some times you just can't see the forest for the trees.
Cobbled together some new leads to the buffer board, powered it up, and still no joy. Started checking fuses and found that number 602 was blown, one of the 38v inputs from the transformer. Installed a new fuse and it blows again. No fireworks or magic smoke, just opens up. Coincidentally, this leg corresponds to the snipped off leg of the 1S1850 rectifier on the board. This modification was the same as the buffer board from the parts carcass, and also the same as was noted by our member Hyperion in post #15 here:
and also an earlier thread of mine:
This was the board as found, with the cut trace corresponding to the cut leg of the rectifier:
and some of the ancillary components of the underside:
I thought I was pretty meticulous in matching what I found, but maybe not. I've pulled the board (kind of hurts to dismantle again, but oh well) and will start going through the work I've done, and start trying figure out the circuit as modified. My comprehension gets challenged past a certain point, so I'm posting a copy of the pertinent part of the schematic :
and as soon as I can, a copy indicating the mods that were on mine in the hopes that someone can suggest what might be the problem.
And Hyperion, if you're reading this, I am a little unclear from your postings regarding the buffer board mods that you encountered whether you rebuilt as you found it or as per the schematic. I figured that since these mods had appeared on multiple boards that they were a factory mod and that it should be rebuilt as found...
That's the story so far- a bit of a setback, but at least other than the buffer board quandary the rest of the unit has kept its magic smoke inside and the bias dialed right in.
A couple of other questions to throw out there...
I am seeing a fairly high Dc offset, around 60 mv, and if that holds once the buffer board is up and running, how concerned should I be? I've read that some pretty high offsets can be the case in designs that don't allow for it to be adjusted, but I'm not sure what to expect from this amp.
Also, regarding the the 15A "midget" fuses in the speaker outputs and the mains input, should the mains be a "normal" blow and the speakers be "fast"? Between the mixture of fuses in this and the parts unit, there seem to be some of the original Toyo fuses, but 2 different kinds. The mains fuse was marked with a number that I found a reference to as "normal," but the speaker fuses have no identification other than voltage, amperage, and brand. Any thoughts on this?
Thanks in advance for any suggestions-
IIRC rebuilt 'as found' apart from the rectifier that looked hacked about, and small changes to enable both diodes of the replacement. I couldn't believe the modification was 'factory' as it looked so badly done, and didn't make sense.
Can't disagree with how they looked, but both of mine and yours, judging from the photos from your thread, appear to be exactly the same, down to the color and gauge of jumper wires and the clear shrink tube on the additional resistors. Seems too much of a coincidence to be a mod that was done in the field. But who knows...
Trying trace out mine, here is the circuit with indications of what was modded:
To my fairly uneducated eye, they have cut the trace from the rectified output of D606 and then fed the un-rectified 38V of one leg of the center-tapped transformer (input 610) directly to C601 (marked C604 on the schematic), R618, and R619, and also jumped the now half-wave rectified output of D606, via an additional 26kohm resistor and 3.3uf capacitor to the base of TR604.
As you mentioned in your thread, this direct jump to TR604 seems to be a compromise to the protection circuit, but it seems odd about bypassing the rectifier to feed the other areas on the board.
It is the fuse for the straight transformer leg that keeps opening, so I will start going through things downstream of this fuse and start trying to make sense of the mono/stereo relay system ...
Okay, some things are starting to become clear thanks to Hyperion's thread that I cited earlier. Going back and fully reading the thread, it struck me in looking at the photo of the seemingly mis-connected tags of the power leads to his buffer board that the un-rectified lead from the transformer tap went to F601 and subsequently to the operational half of the S1850. The other lead, to F602, came from a rectified half of the SG-5T diode, and it is this feed that is jumped past the S1850 diode and on to C601,R618, and R619.
Dug back in the disassembly photos and found that this was also the case on mine.
Have to wonder why they would choose to split the rectification like that?
Of course, when I realized that I had never installed the power leads for the buffer board I just went by the schematic, so it was my mistake in feeding AC into F602 and the bypassed leg of the rectifier. Hopefully this is the problem, and hopefully I haven't done too much damage...
Following Hyperion's lead, I will also install a complete rectifier and return the board to a pre-modification state, with a 470uf cap at C601(4).
Keeping fingers crossed again
More progress, but no success so far.
The 1S1850 (D606) was replaced with a pair of robust 1.5A fast recovery diodes, the cut trace was bridged, and the jumpers were removed. There are still two additional resistors tied into the signal output from the board that I haven't pulled yet. I want to have a little better guess at what they are doing there before removing them...
Powered back up with a nice satisfying snap of the soft-start relay, but still no output. Oh well. I wanted to eliminate issues further down the line so I did a direct feed to the driver boards from the preamp and each channel has speaker output. The left meter isn't responding, but at least I fan isolate the issue at the buffer board.
I pulled the meters to gain access to the buffer board in situ and started tracing voltage readings. The first thing I find is no voltage at F601, half of the rectifier input. The lead from the transformer reads the appropriate 38v, so it appears to be a bad solder joint at pin 609, even though I had routinely re-flowed all of the solder joints on each board. Re-soldered the joint and still no continuity, so figuring it's a bad trace or damaged solder pad so I soldered in a jumper to the fuse bracket. Okay, now there's 38v into the other leg of the rectifier, but still no amplifier output. Trying to think of a fault that would completely mute both channels, I check relay 602, the small reed relay just upstream of the output terminals. It appears the coil is an open circuit, and with voltage applied to the coil there is no continuity through the relay.
That was enough troubleshooting for today, so tomorrow I'll try swapping out the relay for one of the spares from the parts unit. The two originals on this board tested fine when I rebuilt the board, and were not as soot blackened as those on the parts unit but maybe 602 didn't survive the attempts to get this thing up and running. At least the driver and output sections appear to be functioning properly so far and just the buffer board will need some sorting out.
I am not being entirely successful at this point...
Replaced the relay, still no engagement. Out with the board again, check all the traces and values. Thermistor reads correct at room temp and resistance rises as heated. Everything is as per the schematic. Back in- no go. Out again, and I pull transistors 1-5. I had replaced all of the 2sc1364's with ksc945's, so checked them all and made sure that I had in fact bought the center collector 945c version, and those were correct. Back in. No go.
Now I figure that somebody was on to something when they did the mods as I found them. So I pulled the board again and (partially) re-did the mods that were there. Pulling from the new rectifier (D606), I jumped power to the added cap and resistor at the base of TR03. I see that I incorrectly indicated these mods at TR04 on the schematic above.
I had also removed the other resistors that were piggybacked onto the board. These were (I think, but don't have them in front of me) 25k resistors tying the left and right signal outputs to ground at output terminal of the board, and also a wire jumper between the left and right ground pins. I did not add these back in.
With the partial mods reestablished, the board went back in and after a couple of seconds we have a green light and throughput from the amp. Not quite a eureka moment, but I am willing to admit defeat on this and acknowledge that I am in over my depth on trying to understand why the protector circuit does not seem to perform as designed and built. Maybe down the road I can figure it out and attack the board again, but for now I am going to be satisfied with powering up.
Except that there appears to be another anomaly that presented itself while letting it idle to adjust the bias. The outputs are getting steadily hotter as it sits, no load, no input. The initial bias setting was a bit high (in the 190ma range), and I have been inching it back down, replacing the cover, let it stabilize, small tweak, replace the cover...
Even after overshooting to the low side and tweaking my way back up, the heatsink is noticeably warm. That was the end of yesterday's efforts, shutting it down and calling it a day. Did some googling, and now am wondering if the amp is oscillating, which apparently will heat up the outputs with approximately correct bias and no input or load. Yay. Time for some more sleuthing...
At least I can take some consolation that, at least from a broad strokes perspective, the unit does in fact work. Mostly. And can hopefully get fully sorted out. And at least from the minute or two that it I had it playing music, it seems to sound quite good, even through my little test speakers. Now it's time to settle in the bias settings and get it up on the scope...
Thanks for everyone's positive encouragements, and thanks in advance for any insights on its current status.
Hey man these can occur. Steady as you go. BA5K should be cold at idle and even at low power . It might get just warm to room temp after some time it pumps out some power . Lets say halve it's rated power.
But only this. Only just a bit warm , not REALLY WARM OR EVEN CLOSE TO HOT. If you're getting me. Get the picture here?
Separate names with a comma.