|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.131.213.194
Hello Everyone,I've embarked on the rebuild of a nearly-dead SCA35. It will power up with no smoke or angry red tubes but makes horrible howling sounds (through the speakers - quiet with no speakers). In preparing for this voyage, I got a reprint of James Lin's AudioXpress article and have read the Skaggs & Crawford Glass Audio article he references. I've also been on the Roger Stevens site, http://home.netcarrier.com/~rstevens/sca-35.html and had some helpful input from various forums. So I have a general plan on what I want to do but have some questions and need some clarifications (especially if Jim is around).
Here are the phase 1 plans...
1. Change out input connectors. In that process I plan to remove the extra low-level inputs so I'm left with one phone input. In the process the input equalization PEC goes away.
2. Upgrade the power supply. I've ordered Sheldon Stokes' PS board and will use fast recovery diodes will establish a star(ish) ground system. I'll also change the input power to a 3-prong, fused IEC.
3. Circuit/wiring upgrades. I have ordered new circuit boards from John Sullivan (mine are a little crispy) and will use carbon or metal film resistors and modern caps throughout. I'll also require the whole deal 22 teflon hook-up wire.
4. Bypass the stereo/mono, loudness, and filter switches. I'm trying to get a much out of the circuit as possible.
5. Bypass tone controls. I don't want to switch them in and out as Jim and Roger did. I just want them out as I've done with favorable results in my PAS3.
6. Upgrade volume and balance controls to stepped attenuators and maybe upgrade the selector switch. I can live with the cross-talk of the selector switch since I'll have one input live at a time but if there's a sonic improvement, I'd like to go for it -- opinions encouraged here.
The questions start with the removal of the tone controls. Jim had written that the tone control network introduced a 17dB loss in the midrange so when switched them out of the circuit, he added a voltage divider. This makes sense to me but what I can't match up is the loss: I get -15dB for power (proportional to R squared?) I know this is simple so I've got to be missing something -- please for forgive my ignorance, it's been a long time since I've studied this stuff.
Charging forward, in my desire to upgrade the volume and balance pots to stepped attenuators, I have discovered that acquiring anything over 100k Ohms is hard to source. A helpful soul over at diyAudio has pointed me in the right direction but, given the cost and the overall attentuation issue if I'm not better off redesiging this section and taking a more standard path. FYI, the pot values for the volume and balance pots are 250k and 750k, respectively.
I'm confused on how to move forward. On one hand, I think the selector-swith-input-to-power-amp-input circuit can be redesigned using 100k stepped attenuators for volume & balance. I think this is a modified voltage divider arrangement that may simplify the circuit I'm not sure what considerations are important to the input of the 7199 circuit.
On the other, will I get better sound by staying with stock values and add a 17dB (or whatever it is) attenuation between the volume and the balance controls. To further confuse the matter, Jim added a 500k Ohm (not 250k) volume control replacement and Roger added no attenuation at all when taking out the tone controls.
Any help clearing my fog would be greatly appreciated.
Follow Ups:
I rebuilt the phono board to a modified PAS phono circuit, as designed by Joe Curcio. It's relatively simple to do, much more accurate than stock, and is very quiet even with AC on the heaters. Here's a step-by-step that I posted in the Dynaco Doctor forum a while back:
http://db.audioasylum.com/cgi/m.mpl?forum=dynaco&n=8153&highlight=sca-35+phono+Flyquail56&r=&session=Along with UF4007s in the PS, I also used the reverse-recovery spike filter, borrowed from the guys at the Bottlehead Forum. It's not quite tube rectification, but close enough.
I also changed pwr amp boards to ST-35, as recommended by Jim L. and the other folks. Result: my various ST-70s and PAS preamps don't get much use these days :-).
Good luck with your project.
Best regards,
Mike
- http://db.audioasylum.com/cgi/m.mpl?forum=dynaco&n=8153&highlight=sca-35+phono+Flyquail56&r=&session= (Open in New Window)
Interesting alternative. The potential problem with an active RIAA phono stage using feedback is variations in RIAA accuracy due to variations/aging in the tubes. Because the two 12AX7 tubes have a relatively limited overall gain, there isn't much feedback at low frequencies to control the variation. If you look at Lipshitz' JAES article on RIAA EQ, you'll find that the open loop gain is part of the equation to generate the values of the EQ components, so substituting a new tube with a different gain, or aging of the tubes can result in variations. Although passive EQ is not immune from variations, it can be designed to have less variation with tube aging or replacement. OTOH, the feedback design probably has a lower output impedance. Swings and roundabouts, as the British would say.
Hi Jim,
Thanks for your comments. I hope to try the RCA circuit one of these days to see how it compares. A lot of people seem to like the passive EQ.
kurt
try figuring out where the howling sound is comming before you do anything..I have one here that Im working on an sca35 that had the tone controls removed and had a couple multicap rtx caps put in it along with a couple orange drops and while it doesnt sound bad its no scott 299c or sherwood s5000 both which have the gz34 rectifier.It is powerful tho and I measured 22 watts out each channel.The howling may be on the inputs so see if you can short the inputs and make the noise disappear before you rebuild it.If I had one of the amps I would make it into nice st35 with a killer driver stage and that would be a hell of an amp.
Thanks for the thought, Michael, but she's stripped to the bone. I did try shorting the inputs when the unit was still together but to no avail. I didn't take it very far because the unit had been a source for parts so I kluged it to fire it up at all. I'm going to go the path of using the ST-35 design with the RIAA reference phono design. The new circuit boards are ordered along with a new PS board. I'm a little unclear on some of the details that have been referenced in this thread but that the direction. It will be kinda like a kit without instructions. :-) You guys are expert at this stuff. For me, it's unexplored territory.
Are you using the Sovtek 7199 tube ? I tried it on my SCA35, and It had that nasty Hi-pitch squeel. I switch to a american 7199 and had no problem.I bypassed the tone-control, and tried a ST35 board. I preferred the 7199 over the 12DW7. Less highs, but a much more interesting mid-range. Colour sound, yes but it sounds pretty good ! Experiment for yourself and see which board you like.
No comments about your first five items which are basically the same as my modification.Re: tone control bypass, I agree with the others that if you use a straight wire bypass of the tone controls you are better off using the circuit board for the Stereo 35, which has a 7247 dual triode input/driver tube instead of the 7199, which is a pentode/triode tube required to obtain the extra gain lost in the tone control circuit. I actually do say this in my article (p. 51 "If you have permanently bypassed the tone controls your best bet is to substitute the amplifier boards from an ST-35..."). That way you don't have to worry about the amount of attenuation in the resistor network (by the way there is a discrepancy in the parts list vs. the diagram, the diagram shows a 47 kilohm resistor, it should be 43 kilohms which works out to about 16 dB loss, which is close enough). Note that you cannot just stick a 7247 tube in place of the 7199 tube in the SCA-35 board because the pin-outs are not the same and the circuit is also quite different. You will need to get boards for the ST-35 - I just did a search and Triode Electronics sells the clone boards for $39 for the pair. There was also a reconstituted Dynaco which was selling kits and boards for the original Dynaco tube equipment a year or two ago but I'm not sure if they're still around.
The reason for having the attenuation, aside from matching gain with the tone control network so you don't get a blast of sound switching between tone controls in versus out, is that if you do a straight wire bypass the gain in the power amp is too high, which means that you will wind up with the volume control barely off the stops, in the lower end where a small movement will result in a large change in gain, and the tracking between channels is worst. With the attenuator in you'll typically have the volume control near 12 o'clock. This assumes that you're using the original SCA-35 circuit of course.
If you decide to switch to the RCA Manual phono circuit as I did, you need to have the volume plus balance controls provide a load of at least 225 kilohms, so a minimum of 250 kilohms for the volume control is a must. Too low a load will result in blah sound or so I've read. I found my 500 kilohm pot on ebay - 250 to 500 kilohm dual log pots show up there from time to time. If you're going with a dual pot you want log, not linear. If you're using a switched attenuator, Michael Percy Audio stocks attenuator kits, pots and switched attenuators of 250 kilohms but they are pricy. I believe Welborne Labs does also. Incidentally, unless you're planning on using the balance control to any extent it's probably OK to leave the stock pot, according to the schematic, in the center position it should be equivalent to a single contact and is pretty much out of the circuit. If you're substituting another lower impedance balance control I would definitely have it after the volume control so that at reasonable volume settings the phono circuit isn't seeing the balance control as a significant influence on the load.
Incidentally, Roger Stevens messes around with the positive feedback resistor R33 in the SCA-35 power amp design. I do NOT recommend this, because that resistor R33 is also seen in the forward direction as being in parallel with the phase inverter resistor R34 on the cathode of the 7199 triode section. In a classic cathodyne phase inverter as used in the SCA-35 (and also ST-35) design but WITHOUT positive feedback, the plate and cathode resistors MUST be equal, in order that the two output tubes receive the same amplitude signal (but in inverse phase). However you will note that in both the Dyna designs the cathode and plate resistors have DIFFERENT values. In the SCA-35 these are R32 and R34, which are 22k and 27k respectively. This is because the positive feedback resistor R33 is in parallel with R34, so that the parallel combination of R33 (120 k) and R34 (27k) is very close to R32 (22k). Stevens changes the value of R33 WITHOUT altering R34, which not only changes the feedback, but results in one output tube getting a different amplitude signal from the other, resulting in a distorted signal at the output. In order to PROPERLY play with the feedback, R34 needs to be changed each time R33 is changed so that the parallel combination of the two still works out to 22 kilohms. This is one of the subtleties of the Dyna design that many people miss. By the way, the ST-35 circuit ALSO uses positive feedback in its design. In fact there is a version of the ST-35 circuit on the internet which completely omits the positive feedback resistor, but then leaves the cathode and plate resistors on the phase inverter/driver tube different rather than making them the same, BECAUSE THEY DIDN'T UNDERSTAND THE DESIGN! Distortion by design!
James,One thing I forgot to ask you...
Where did you locate the current limiter in the circuit? And if your remember which one you used I'd be interested.
Thanks again. KB
I desoldered the wire going from the fuseholder to the power transformer and inserted the current limiter in line between the two. I used an inrush current limiter rated at 2 amps with cold resistance of about 120 ohms. It's available in the Mouser catalog, GE part number CL90, at $1.72 and there are probably equivalents from other manufacturers.
This is what JimL reffered to see link.
Tazsmonn "The Cat Herder".
W66th.St./Richfield/Minneapolis/MN/USA
Cats are good room treatments, you should have 2 Or 3.
My detracters have never heard my system.
Jim,I like the RCA phono circuit. It's BETTER than Dyna's, but the HIGH O/P impedance liability is present in both circuits. The ZVN0545A source follower tweak is an easy and EFFECTIVE solution to the drive issue in both cases.
The basic RCA circuit can be tweaked in other ways too. Converting the 2nd gain block from cathode to grid leak bias improves bass extension. Connect the 2nd triode's cathode to ground and eliminate the RC bias network. Replace the 680 KOhm grid leak resistor with 20 MegOhms. A change to a value in the EQ network has been posted here on AA that brings things closer to the RIAA ideal.
Eli D.
Eli,I like the RCA circuit for this particular application because it's simple, classic, has much more accurate RIAA EQ than the original, and uses the same tubes as the stock Dyna, but agree that high output impedance is its Achillies heel. Not familiar the ZVN0545A source follower tweak but it sounds interesting. Only problem is that space is very limited (open board space measures appoximately 3" x 3 1/2" and the selector switch overhangs about 1/2 of it) in the SCA-35 - the basic RCA circuit barely fits without additions, and in fact I had to omit the first stage RC decoupling cap in the power supply due to space limitations, and also reduce the output capacitor to a mere 0.1 uf for the same reason. Squeezing in another output cap for the source follower would probably be very difficult if not impossible, but it sounds like an excellent idea in an application with a bit more space.
Changing the 22 kilohm resistor in the original circuit to 24 kilohms improves the RIAA EQ, and I have so posted in the past - the accuracy improves from approximately plus or minus 0.4 dB to plus or minus 0.2 dB, which shows I think that the RCA engineers knew what they were doing. They did as well as could be expected using 10% resistor values. Both are much more accurate than the stock Dyna circuit, which in my sample measured plus or minus 2 dB! Similar results were reported in Hi Fidelity and Audio in reviews of new SCA-35 so it's not just due to aging components.
The 680 kilohm resistor is not only a grid leak resistor but also part of the RIAA EQ, changing that will alter the RIAA accuracy very slightly in the highs but significantly in the lows - the nominal low frequency rolloff will go from 50 Hz to 30 Hz. In theory the values for the cathode bias cap should result in a roll-off of the second stage around 2-3 Hz, which I believe is low enough, considering that we are dealing with an integrated amp not likely to be used with speakers that can get anywhere close to 20 Hz. Not to mention that the low frequency response is further limited in my version by the small output cap. Of course I can't answer for the subjective impression of the changes you suggest. I do wonder though whether the improvement in bass extension is due to the change in RIAA EQ rather than change in the bias method.
The ZVN0545A is an enhancement MOSFET in an E-Line (TO92) case. It gets DC coupled to the 2nd triode's plate. An additional coupling cap. is not needed. Use a physically small 1/4 W. Carbon comp. part as the gate stopper. Setting the FET up is described in MOSFET Follies. Examination of the ZVN0545A data sheet will show SMALL capacitances, that even a 'X7 section can drive well.
Eli D.
Any suggestion as to the follower resistor value? That's not specified in the MOSFET Follies, although a 220 ohm gate stopper IS suggested. As a very rough quick and dirty calculation, the plate voltage in my SCA-35 phono stage runs around 200 volts with a 260 volt B+, so a 100 kilohm follower resistor should give a bit less than 2 mA through the MOSFET, which will result in a resistor dissipation of 0.4 watts (use a 2 watt resistor) and a MOSFET dissipation of 0.12 watts which should be safe. Of course I tend to design conservatively. Not a great deal of current drive but certainly better than the 12AX7.
I'm yet more seat of the pants and ignore the FET's internal resistance. 260 V./150 KOhms = 1.7 mA. That's SAFE. A 1 W. part will do, as long as air flow is decent and air flow is also needed for the FET.Keen's remark about the little FET being a good replacement for a 'X7 CF hits the target. The FET's transconductance (gfs) is nearly proportional to Id. If gfs is about 1.7 mA./V., the O/P impedance is under 1 KOhm. That's certainly adequate to drive a 50 KOhm pot. on the same chassis.
BTW, 510 Ohms or 1 K is what I'd use as the stopper value. Caution on my part.
Eli D.
I would 2nd Ed advise on the 12DW7 and add a bias resister for each 6BQ5. This improves the fine detail of recordings.
Tazsmonn "The Cat Herder".
W66th.St./Richfield/Minneapolis/MN/USA
Cats are good room treatments, you should have 2 Or 3.
My detracters have never heard my system.
Dude,The only reason to use 7199s is the losses the tone control circuitry introduce. If you "axe" all the signal processing stuff, switch over to the ST35 small signal circuitry that uses 1 12DW7 in each channel.
You can avoid a LOT of hassles by a employing a separate 50 KOhm PEC hot molded Carbon pot. in each channel for control of volume and balance.
The OEM phono section circuitry needs a HIGH impedance load. You resolve that issue by buffering each 2nd 'X7 section with a DC coupled ZVN0545A source follower, as described in MOSFET Follies.
Speaking of the phono section, construct a DC supply for those 'X7 heaters.
How much are you willing to spend on a replacement source selector switch? $21.52 will get you Mouser stock # 690-C1D0212N-A, which is an ElectroSwitch 2 deck part that can control up to (sic) 12 sources. $18.09 is the cost of Mouser stock # 690-C1D0206N-A, which is an ElectroSwitch 1 deck/2 pole part that can control up to 6 sources.
Eli D.
This post is made possible by the generous support of people like you and our sponsors: