|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.230.67.50
...I can merely substitute? (Both cascaded 'N7s are SRPPs in the frontend of my ASL AQ1006(845).)Or can I simply replace the 100K input-load resistor with a high-quality 100K pot?
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
Follow Ups:
Jeffrey, after reviewing the schematic, I see that the front end isn't exactly what you think. True SRPP stages have roughly half the gain compared to the same tube type configured in common cathode. SRPP also does not have a bypass on Rk (the lower cathode resistor). To derive some approximate numbers, I Spice modeled the front end of your amplifier. Here are the results:Factory Configuration:
1st Stage VB1 Gain = +23 dB (voltage gain 14)
2nd Stage VB2 Gain = +22 dB (voltage gain 12.5)
Total Front End Gain: +45 dBTrue SRPP:
1st Stage VB1 Gain with C2 Removed: +17.5 dB
2nd Stage VB2 Gain with C4 Removed: +17.5 dB
Total Front End Gain: +35 dBIn other words, if you remove C2 and C4, amplifier gain will decrease by approximately 10 dB. Spice shows frequency response relatively unaffected by this, but of course, there may be "atmospheric" changes that only you can judge. I personally don't configure SRPP stages with bypass caps, so speaking for myself, this would be a simple decision. I sugest you remove those two caps, listen for a few days, and report back. :)
I've also been thinking about a frontend pot. Obviously adding one wouldn't reduce the amplifier noise passed to the speakers, but it sure would reduce, to the extent it's turned down, the preamp linestage noise being amplified by the amp.Also, I recall vaguely that using the 4- and 8-Ohm output taps for common and positive, respectively, gets one 2 Ohms; is that correct? If so, the Voltage gain thru those taps would be even less than thru the 4-Ohm taps. I guess I'll measure that.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
Using the wrong taps will reflect a mismatch back to the 845. Sound quality will almost certainly be compromised. It's also impossible to know how the transformer will perform (frequency response, ringing, etc.) when the lead intended by the manufacturer to be used a common is left open.I'll add that the SRPP modification is probably the only simple means to reduce gain by an appropriate amount without degrading S/N. A stepdown transformer at the input will reduce linestage noise, but high sensitivity speakers will still reproduce internal noise of the amplifier's front end. On the other hand, bypassing the first stage altogether will reduce amplifier gain too much (23 dB). I think you would need a high level source to compensate for such a large reduction.
I'll start with the 2nd, since that will reduce preamp and poweramp-1st-gain-stage noise, while removing the 1st cap will (presumably) reduce only preamp noise.Again, thx for your help. I'll let you know what happens.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
It's the easiest way of reducing gain and way better than a pot IMO. It will change the sound a little and you may or may not like it. You should also know that removing the bypass cap will also degrade PSRR by almost 50% and increase the OP impedance by a similar amount but neither should be a factor in your circuit.Actually the first stage is not an SRPP at all since the drive is taken from the plate of the bottom tube. It's more like a symmetrical CCS grounded cathode stage but the 1K plate R provides a very poor load. I'm not a fan of this configuration because it forces the use of the same value plate R as the cathode R and this compromise leads to a quiescent current and a plate R that are both too low for good linearity. It also (normally) forces you to bypass the cathode R for good gain and low OP impedance. It would be simple to improve by adding a few components.
For the solution to the specific problem you seek TK's suggestion is the best. For other benefits with simple mods consider also 1) taking the drive from the cathode of the top triode and 2) increasing the value of all cathode and plate Rs to between 1K5 and 2K2. Here, running the first stage from the same B+ as the driver would help to keep quiescent current up. The HC voltage of most 6sn7s is 200V so should not be a problem. Also, what you'll lose in higher PS noise you'll gain in better rejection. These mods are simple and worthwhile evaluating seperately. I'd be interested in your feedback should you try any of them.FWIW, both stages are not very linear but the driver particularly so due to the high voltages it must swing (which would be helped considerably by mod 2), however distortion will be mostly 2ndH and I've found that this is actually desirable in SE driver stages since it goes a long way towards cancelling some of the distortion generated in the OP stage since it's opposite in phase.
.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
The problem with throwing away gain, is that it increases noise. It's best to have a sensible gain-structure in place to start with. This of course requires significant changes to the amp.If this isn't an option, then some sort of divider, or step-down transformer at the input maybe.
Possibly, you could move the signal feed straight to the second stage.
You know, 27dB gain doesn't seem excessive. If one starts with a 1V RMS input and amplifies it to 14V (~24.5 W @ 8 ohms), that's 23 dB. Your linestage must be averaging quite a bit more than that. In any event, ya gotta do what ya gotta do. :)
All else being equal, and assuming S/N is not an issue, I would recommend a fixed resistive divider at the front end. That's the only modification that will not require considerable other circuit changes, and which is likely to have the least sonic impact on the amplifier.
.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
Jeff, can I ask the reason for wanting to do such a thing because the answer may be different depending on your goal.
...and discovered just how much the amps amplified and the speakers made more audible the noise of the linestage's tuberush.I think this particular system could do with at least 10 to as much as 20dB less gain than its rated 27dB. Previously I had my 12dB-gain Music Reference EM7 amps connected and the system was virtually dead quiet, noise being audible only with one's ear virtually inside the horn.
If as Triode_Kingdom wrote, a fixed Voltage-divider network is appropriate, could it be installed after the 2nd 6SN7 gain stage, say after the 2nd coupling cap, thereby quieting the fontend's inherent noise?
I have copies of the schematic I'll send to anyone who e-mails me at jeffreybehr(at)cox(dot)net.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
Hi ,
Sounds like a job for a TVC/AVC . Maybe it's time to lose the linestage ?cheers
I've already emailed you for a schematic, looks like you can do away with a complete gain stage which would be even better than a pot or divider. Perhaps you could install another input jack coupled directly to the second stage but this will depend on other factors such as input impedance etc. The fewer stages the better and matching total gain to that required is always best.
...gain stage sounds quite interesting.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
First, it's not all that straight forward to eliminate the first stage because the second is referenced to neg 133V instead of ground. The easiest option here is to IT couple from a new input to the second stage but the downside is that you need very good iron and a place to put the trannies. I may be in the minority here but it wouldn’t be my first preference.Opt 2 is to Ref stage 2 to ground but then you'll be marginal on linear swing to drive the 845. This can be easiest fixed by increasing the B+ to stage 2 (not a lot of headroom here though with your supply, maybe 20V) in tandem with a favourite of mine, changing the plate R of the stage from a 1K to an inexpensive choke measuring around the same resistance. If you have difficulty finding one even a Hammond 154E will do but because it measures 1.666K you'd best change the cathode R to 1K68 as well.
There is a third option which is a much bigger job but can yield great gains. You could convert the driver to a White CF which has slightly less than unity gain and keep the first stage. Advantages are extremely low OP impedance to drive the 845 and the ability to drive the grid positive yielding around 40W for greater headroom. You'll also eliminate one coupling cap ... always a good thing. The downside is having to increase the neg supply voltage to the driver to at least -360V. You'd also need to reduce the B+ to say 120V or so. Well worth the trouble IMHO.
Cheers,
Naz
-to eliminate a stage of gain.If you are really thinking of going with speakers that efficient that might be the way to go (if you don't mind working on the circuit).
It is also possible to rebias the 6SN7s by changing the cathode resistor to a larger value. You could reduce the gain by about 8-10 db this way without suffering too much sound-wise, but eliminating a stage of gain is *always* going to work the best, if you can.
...not a designer. :-)I hope someone can come up with the exact method to add another inputjack that bypasses the 1st gain stage.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
...are pinout-compatible with 6SN7, but gain is only moderately lower, and the plate resistance is considerably lower in both; so circuit modifications might be needed. You can look them up on duncanamps.com.
...that you also have to pay attention to the filament draw. In the 6BL7 it is also much higher than the 6SN7, so you would have to be sure your power transformer could handle it.If you are going to consider transformers, I suppose you could replace the coupling capacitor to the output with an IT (in each channel) connected as a step-down. The voltage ratio equals the turns ratio, while the impedance goes like the square of the voltage, so this would be beneficial from the current drive standpoint as well.
Be very wary with the 6BX7 since the filament takes double the current of a 6SN7. It will aslo give a cleaner bass due to the low Rp.
I would suggest an alternative solution in that you can fit a Lundhall input trannie: this can be configured to give 1:1 or 2:1 or even 4:1 without sacrificing the cleanliness or the dynamic range of the incoming signal. This resolves the problem of excessive gain when using horns.
I use the nickel core type for this purpose.
Hope this helps.
John R
.
-------------------------------------------------------
Tin-eared audiofool and obsessed landscape fotografer.
http://community.webshots.com/user/jeffreybehr
This post is made possible by the generous support of people like you and our sponsors: