Fisher 400--Locating under-chassis Caps

[Dave451]

I am rehabbing another Fisher 400 and installing the usual mods (IBAM, screen grid resistors, 10 ohm cathode resistors, change from 330K to 200K grid resistors with 0.1 uF couplers). I'm also replacing the electrolytic caps, including both cans with axials under the chassis.

The original elements of the can caps are centrally located (being in the same housing) and the filtered voltages led from the + terminals of the caps by wires to where they are used (output tube B+/screens, PI tube plates, etc) with the grounds centrally located around the can lugs.

Here's are my questions: is there any reason that individual axial replacement caps can't be placed where the voltage is actually used in other parts of the chassis and grounded there, instead of clustering all the axial replacements in the location where the original cans resided? Is this going to produce some sort of 'ground loop purgatory,' with hum/buzz or is it feasible? Does the central location of the ground points around the cans act as a sort of 'star ground' or can the caps and their grounds be dispersed without noise?

Thanks
Dave

 

[dcgillespie]

Dave -- For the Doubler and Screen cap sections, they should definitely be grounded where the original can cap was. As you move into the remainder to the cap sections (down the B+ stream), they can be located near the circuits they serve, since their function is primarily to decouple only, where as the first two sections decouple and also pass significant ripple current as well. When a B+ cap is not passing any ripple current (of any significance), its ground location is far less important relative to the formation of ground loops -- other than the fact that the ground point be a good, solid ground connection.

Dave

 

[larryderouin]

Dave. I've done what you suggested on a 481A(Custom Electra VII amp), and a 49A(Custom Electra VIII Hybrid Amp) with no problems and no ground loops. Fortunately none were the V.D. or Screen sections. Both are still running very well after 5 and 9 years. You'll probably find that moving them will involve possibly drilling out rivets to install nuts and bolts to attach terminal strips, provided there is room in the particular area. Mock it up 1st and see if you have the room in the areas needed for the caps. Some areas on the 400 can be tight. If the area is next to the outer walls, and there is a hole already there, use it and hang the terminal strip and caps from the wall. A panhead screw head shows up less than a regular round head.
Larry

 

[Dave451]

Thanks, guys, for the help. Dave, the explanation makes a lot of sense and always good to understand the underlying reasons why something is practical or not. Larry, I'm in the mock up stage now to see what looks practical. One consideration is that I need room to mount my IBAM board, so I'm trying to be efficient with space in that general area of the chassis, perhaps by spreading out the axials. In the last 400 I did (I did a thread on it), I replaced all the cans with Hayseed Hamfest cans, but I want to try the all-under-chassis approach for this one. I'll send some pictures shortly after I work with it a bit.
Thanks again.
Dave

 

[Dave451]

Working away on the 400. Another question: I'm replacing the 0.01 uF "line to ground" 600V tubular cap with an X1/Y2 safety cap (300/400 VAC). In the 400, there's another 0.01 uF 600VDC cap from the secondary to the chassis. Should this be replaced in kind or with another safety disc cap? If so, what rating?
Thanks!
Dave

 

[larryderouin]

Regular .01 uF 600vdc cap. It's NOT a Death cap. IIRC it's there as as a snubber to damp out switching noise from the diodes.

 

[Dave451]

Thought so, but wasn't sure--thanks Larry! I'll replace with 0.01 uF 630VDC. Pictures coming soon!

 

[Dave451]

Work is well underway, with all the electrolytics in place as per the pictures below. The 100 ohm screen resistors (with their bus), the 10 ohm cathode resistors, 200K ohm grid resistors, and the 0.1 uF 630VDC IC MPW output tube couplers are all also in place. I put in a CL-80, visible in a couple of the shots (had to replace the fuse holder as well, believe it or not) and reworked the bias supply with new 1000 uF 35VDC caps, leaving the silicon rectifier bridge in place. As per the above posts, I replaced both of the 0.01 uF 600VDC tubular caps around the power transformer: the PT primary line-to-ground ("death") cap with an 0.01 uF X1/Y2 rated disc (orange) you can see, and the other (in the PT secondary) with a straight 0.01 uF 630VDC film tubular. I also replaced the two 0.01 uF disc couplers in the pre-amp 12AX7 stages with film caps.

For this rehab, I am using Drew Boice' IBAM design (I had the parts on hand) after using Dave Gillespie's design for all prior Fisher work. The picture shown are taken after all the above work, but prior to the IBAM installation. The nice space where I removed the old 2x1000uF bias cap will be the home of the IBAM. C42 (100 uF), R34 (2.2K) and R35 (1K) are unused and are all removed, per Drew's recommendations.

So, here are some shots of the tubular electrolytic installations. I used axial tubular replacement caps instead of radials--personal preference. By the way, I replaced all the cans EXCEPT for C56, the hot can (cardboard covered) 100 uF 300VDC doubler cap, which (believe it or not) easily passed the max working voltage leakage test on my Sprague TO-5 checker. The other two cans (C52 and C57) had at least one section that failed the leakage test and C57 was visibly leaking electrolyte.

Here's the bias supply with the much smaller modern 1000uF 35VDC caps installed and the silicon bias rectifier bridge re-wired in a more compact configuration. You can see the CL-80 installed in a PT primary leg. As per above, the 100uF cap and two resistors are not used and have been removed.




Below are the replacements for the C57 can (200uf 300V, 2X50uF 250V). I used 47 uF to replace the 50 uF sections, as these are well within the tolerances of the original units. As you can see, the 220 uF 300V replacement for the C section is a pretty good size cap. The 220 uf and one of the 47 uF caps are grounded where the 25 ohm speaker resistors are grounded. The other 47uF 250V is grounded to a can take of the original C57. Hopefully I'll get away with this without buzz/hum.

You can also see where I mounted the replacement for C52B (40 uF 400VDC), which decouples the tone control 12AX7 plate supply, the (mis-) labeled 390V supply (I think the 320V and 390V B+ points are switched on the Fisher 400 late model schematic). I used 47 uF caps to replace all the 40 uF sections of C52. This one is mounted directly to the "390V" B+ tie point for the tubes, with the minus on a nearby chassis ground point.



Below are the other two replacements for C52C and C52D (40uF 450VDC and 40 uF 500VDC respectively). I used 47uF replacements of equivalent working voltage. In this view, you can see the orange replacement for the 'death cap' as well. The D section 500VDC replacement is for the screen supply and you can see the positive end connected to the end of the screen supply bus for the output tubes (more on that below). The 450V axial next to it is the replacement for C52C (40 uF 450VDC), which decouples the phase inverter/PI plate supply, the (mis-) labeled 320V supply on the schematic. These two are grounded to the original C52 can tabs.



Here is another view of the C52 C and D caps showing where I added a two-point tie strip in addition to the single-point tie already in place for the B+ resistor string of R40 (1.2K 1W), R39 (2.2K 1W), and R38 (22K 1/2 watt). The purple and orange wires visible connected take the respective B+ voltage to the decoupling electrolytics that are remotely mounted (C52B above and C52A, below).



Here's the replacement for C52A (20uf 300V), mounted remotely with the + connected to the single plate tie point for the 12AX7 pre-amp tubes and the (-) connected to a nearby grounded tie point. I used a 22 uF 350VDC cap as the replacement for C52A. Also shown are the two film couplers added to replace the 0.01 uF discs in the pre-amp tubes, as recommended by Larry. Note that the replacement for R33 (100K 1/2W) is installed here, just below (and across) the electrolytic and grounded to the same tie point.

 

[Dave451]

Here are couple of shots of the work around the 7868 output tubes, showing the 10 ohm cathode resistors 100 ohm screen resistors (with bus), new 1K and 200K grid resistors, new 0.1 uF coupling caps. You have to use some kind of bus for the 7868 screen resistors because there are no unused pins on the 7868 like there are on the 7591 for you to use as a tie point for screen wiring (the tube sheet says DON'T use pin 8, even though nothing is connected to it below--internal connection). There are two single-point tie strips to which each end of the screen bus is connected.

Here are the 7868 sockets with the various resistors installed and the coupling caps visible. The 1.2K 7W (R41) B+/screen supply dropping resistor is moved from one of the 7868 screen pins to a single point tie strip I added at the far end of the screen bus (with red insulation). The right most end of this resistor is connected to the 405V B+ point (which also is the take off for the 7868 plate supply going to the output transformers and, through them, to the 7868 plates) and the left end is connected to the bus providing the screen voltage to the output tubes.



Here is a detail shot of how I added a two-point tie strip to connect the bias supply voltages from the IBAM to the 200K grid resistors (IBAM not yet connected). The added tie strip is the one with the nut holding it down. I had to loosen the output transformers to install the screw and nut for the tie strip, definitely a PITA but do-able if you treat the transformers carefully. The laminates didn't end up touching the screw head (used a #4 pan head screen for lower profile).



So, going to hook up the IBAM and then do some voltage checking at lower voltages with no tubes using the Variac before going further. More to come! Hopefully the Magic Smoke stays in!
Dave

 

[RS Steve]

Tidy work Dave, very nice!

 

[Dave451]

Thanks, Steve! After a triple check of all the wiring, I removed all the tubes and ran some lower-voltage checks using the Variac at all the B+ and bias points. Everything is acting right so far, so I'm going to put all the small tubes in and try it without the output tubes.

 

[bhamham]

Really nice work, Dave451!

Very impressed, a clever way to not have to deal with those troublesome cans.

 

[Dave451]

Thanks, Bruce! Going with good quality axials 'under the hood' can be more economical than buying replacement cans, especially if you're replacing more than one, but I wanted to see whether I could spread them out more instead of clustering them all around the location of the original caps. It does take a bit more bench time, though to drill out rivets, and install the needed tie strips without drilling new holes, but it can be done! It's not rocket science and not even an original idea, but I wanted to try it out and see if I could do it neatly.

I installed the Drew Boice IBAM. Pictures are below. I put the board on a stand off attached to the rivet hole where the C60 1000uFx2 cap lived before and it fit nicely in the space. The 400 chassis is plenty deep, so I could mount it above the other components and not bump into the bottom plate. You can also see how I ran the wires to the tie strips for the 200K grid resistors for the 7868's.

I pulled all the tubes and brought it up on the Variac. Everything behaved well and I tested up to 300VDC on the B+. The bias supply responded well and the trim pots gave me about a 2X swing in negative voltage to the output tube grids (10K trim pots, 12K resistors, and 33uf 50V caps are the parts list). No snapping, crackling, or magic smoke, so I installed the small tubes and tried again. Same results, so I put the output tubes in with the bias set at max negative and brought it up on the Variac again.

All was well, so I set the line voltage at 120VAC (about 118VAC exit the CL-80) and biased the tubes to 32-33 mA no-signal with no problems at all. The adjustment was smooth and the B+ voltages came down as the tubes loaded up. Here's what I got for voltages (schematic values in parentheses):

B+ = 411 (405); Screen supply = 366 (365); "390V" supply = 336 (390); "320V" supply = 297 (320); 210V supply = 210 (210); Bias -25.9 VDC

The output tube plates were all at 407 VDC (405), the screens were 366VDC (365), the grid voltages were all -16.0 to -16.2 VDC, and the cathodes were 330-337 mV, indicating cathode currents set at 33-34 mA, or 12.2W plate dissipation for the time being.

So, the 390 and 320 volt supplies are a bit low, even though the voltages on either side of them on the B+ string are spot on. The resistors checked within tolerance, so the differences must be current draw in the individual stages. Should I worry?

EDIT: I checked the schematic for the early model 400, for which the power supply is identical and it lists these take offs as 320V and 286V, compared to the 336V and 297V I measured, respectively, so I think we're in OK shape here. Looks like the late model schematic error is labeling the next to last take off in the upper string as "390V" instead of reversing the 320 and 390, as I had thought. Dave

The best news is that the unit plays and sounds very well--no 120 Hz at all in Aux or Phono/Tape with no signal and volume max. The inputs are not shorted and I'm getting a very little bit of 60 Hz, but the tube shields and the covers are off and it virtually disappears when you touch the chassis, so experience leads me to believe that will go away when she's buttoned back up. All the Fishers I've worked on are very quiet that way.

The tuner is working well with good behavior tuned through the signals, evident good sensitivity (even with a clip lead antenna in the basement!), and the tuning eye is working well (I am going to put in the 47K resistor mod to reduce the 'dancing' a little). There is no distortion I can detect by ear in the FM stereo signal and MPX noise is quite low. I did replace the ratio detector 8uF cap and the MPX 1uF 350VDC cap, so I'll check the front end alignment but everything seems pretty good at this point!

I'm going to clean her up and get things adjusted (AC balance in particular) and will report back, but it's all good so far.
Dave

IBAM in it's new home:



IBAM wiring to output tube 200K resistors:

 

[larryderouin]

Yeah, FISHER Screwed up the voltages on the late model schematic from C52B. Should be about 286v according to the 10001 and 20001-29999 manuals. If your voltages are falling in line with the older modeling you're ok. Make the change on the late schematic to 286v as nominal and go for it. So far it looks really good. And ifit's working as designed on the variac at lower voltages, throw in the outputs, close up the case, put the shields on and play the hell out of it. Nice job Dave!

 

[Dave451]

UPDATE: I finished the work on the 400 except for the 47K resistor for the "dancing eye" issue. All is well so far. Adjusted for 36 mA cathode current on the output tubes and power tested the amp. The 7868's on the right side are a little weak, left side fine. R power is 18W RMS and L power is 23.8 WRMS. Recall that this is a late model 400 with the 10K output transformers, so I think around 23W is the max at clipping (as it was for the 400 I modified in my other thread with Dave G's mods). Please correct me if I'm wrong on the max output power for the late-model 400's.

Will need to replace the weak tubes, but sound is good and am moving the 400 up to the listening room for listening and tuner testing. I also replaced the marginal 6HA5 front end RF pentode with a NOS Amperex tube, so hoping sensitivity will be good. I plan to align the tuner and MPX65 after some listening.

 

[larryderouin]

Dave; you're correct in that 23watts is max output power for stock late model 400. If you find that you have an imbalance, and everything else is working as designed, back off the left side to 18w until such time as you can get better tubes.

 

[Dave451]

Thanks, Larry—will do. Will be talking to Jim McShane about some EH 7868’s I guess. No spares on hand and will just have to adapt to the big pins.

 

[Dave451]

UPDATE: "Slipping the Noose!"

I purchased a matched quad of EH 7868's from Jim McShane and carefully installed them in this 400 (used a solder pick to gently enlarge the tube socket pins before carefully inserting the EH tubes, which have the large magnoval size pins). I biased each tube to 35mA cathode current using the IBAM and checked all the voltages, which were nominal. The grid voltages for the EH tubes were running between -18.5 and -19.5 VDC; about a volt higher than the old tubes.

Unfortunately, the unit still underperformed, particularly with both channels driven at 1 kHz into 8 ohms (17W on right and 20W on left) despite fiddling with the AC balance controls for minimum distortion. I puzzled over this for a while THEN, I recalled Dave Gillespie's work on the Fisher "noose" in the PI/driver circuit (particularly what he wrote in the "Fisher 400 Service Bulletin" thread). I had installed a new set of Russian Sovtek 12AX7LPS tubes for the PI's and thought 'uh oh.' In a nutshell, Dave informed that the 150K R115/116 'noose' resistor between the PI plate supply and cathode 'strangled' the driver and caused early clipping, particularly with new-production Russian tubes which don't operate as well as the older tubes in the near-cut off zone that the 'noose' forces on Fisher receivers, particularly the 400.

I substituted some old production 12AX7 pulls and the power immediately went up! I disconnected the 'noose' resistors and re-set the phase inverters with the AC balance trim pot (big difference in the setting after removing the resistors) and was able to get a little over 22W in the right channel and 23W in the left channel even with both channels driven and the Russian PI tubes installed; right at full power level for the late-model 400.

By the way, both resistors checked low (120K and 125K), so the strangulation was probably even worse than design. Between that and the higher negative bias voltage for the EH tubes, the PI/Driver was probably laboring even more.

After moving the 400 back to the listening room, I'm also noticing less rolled-off highs with no noose and the new EH output tubes. Very nice! I may be imagining the sonic improvement, but it sure seems to be there. Anyone else note this after noose removal?

Note to Self: Always remember what Dave G says about Fisher (or anything else!).

Dave
PS Picture of the culprits below! (not sure why one was 5% and one was 10%--go figure)

 

[larryderouin]

Nicely Done Dave. I've noticed that also on my 400's. Even with the noose gone I've noticed the Russian tubes still have trouble keeping up when pushed hard. Some of the EH, Sovtek and Tungsols I have are 5 to 8 years old tho. Maybe the newer ones are better? So I use Fisher Tele's in P.I. posit on all my receivers that use 12ax7's.

 

[gadget73]

Were you having any issues with the output of the phase inverter collapsing and then resuming at higher power levels? That was the issue I ran into using Sovtek tubes in a phase inverter. Different design and a different amp but similar low current operation. Old tubes made it behave properly. It was actually the first stage that was dying, not the inverter itself so I lost grid drive on both tubes.