• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Audible Illusions Modulus II (C) Schematic?

grindfix

grindfix

Full time hobby
Subscriber
The actual schematic seems to be well protected by AI. I wonder if someone went through the effort of drawing one? Is there a known solution to lower filament voltage? Mine runs tubes at 6.67VDC. Voltage is regulated but why were they designed to run tubes so hot?

Also, another issue.
This unit has been recapped in 2020 but wasn't really used so doesn't have many hours on it.
Filament rectifier diodes were replaced with a tabletop bridge with the heat sink on its back.
Tubes are 6922EH from McShane. Those were briefly used in another preamp and have several hours more on them than this preamp itself after the recap but I'd still consider them near new.
The frequency response drops fairly quickly above 5KHz which is also supported by the rolled-off leading edge in the square wave display. If I have to guess, those tubes have less than 20 hours on them, and half of that was in the VTA PH-16. If this Modulus ate them this quickly, something must be done before I start looking for new tubes for it. Any suggestions are appreciated.
 
Last edited:
Register to hide this ad
Not having any replies I was prepared to draw out the schematic of the filament circuit this evening. Turns out, no regulation. Standard Rectifier in bridge orientation -> CRC -> Output.
Replacement of the original 1Ohm resistor with 1.33Ohm sets heaters at 6.35VDC.

Now I have to figure out the frequency response issue. I'll take a chance at drawing the audio circuit schematic and will post it here for review.
 
This seems to be an often asked question and no one seems to have the schematic. No doubt your efforts will be appreciated by many.
 
Feeling some pressure, Kevin! :)

I have to add a note, hoping that one particular member will see this. I don't remember this member's ID and can't find the post where he/she said that he/she pulls Phono tubes out and uses an external Phono preamp because heaters are regulated. NO! They are NOT regulated. This causes filament voltage for AUX section tubes VERY HIGH! This takes 0.6A draw out of the circuit. See my post above.

Now, hopefully, I got this one without mistakes. I have never seen this topology.
They used odd 1% resistor values that match my measurements pretty closely.
Any opinions and/or suggestions are welcome!



Audible Illusions Modulus II (C) AUX - Scheme-it.jpg
 
That 100k resistor at the input will create a low pass filter. The volume pot makes one too, but the response is variable. Wind the volume full up and see if its still got the roll-off. If it does, bridge the 100k with a small value cap to compensate for the loss. Probably going to be something under 100pf, if you have a variable cap out of a junk tuner that might be exactly the right device to dial it in. Tweak it so the square waves are properly flat and install that value cap. Or round up if needed, since the volume pot will add some extra roll-off and slightly over-compensating the fixed resistor will average out with the loss in the volume pot.


or just dump the resistor if your source can work into a 100k load and it doesn't end up producing excessive gain. The 100k resistor in series with the pot will get you 50% of the signal from the source to the grid of the 6922

related to that, if it does make too much gain, ditch the cathode bypass cap. That will introduce local feedback and reduce the gain a fair bit.
 
I've had a Mod3A since buying it new 25 years ago, it's on 24/7, and I have found Pearl Tube Coolers on the tubes increase their lifespan a great deal...no affiliation with Pearl, just a happy user...the Pearls also let me use other tubes than the 6h23neb or 6922EH and not have them go microphonic in a month.
 
All the square wave testing was done with the volume full up.
I understand all the suggestions but... I believe these preamps are known to sound good but hard on tubes. This one did not sound impressive to me, that's why it was on the shelf. I initially replaced filament rectifiers with single diodes and reinstalled the copper heatsink wings back on them. Those were still running way too hot so I never used the preamp again. Now, I'm trying to catch up on unfinished projects and got the preamp on the bench this weekend with the idea of replacing single diodes with the bridge on the heatsink. Once that was done with the success, I started testing it further.

I'm sure I can take the above-mentioned steps to make it closer to a flat response, but why is it performing so poorly and equally in both channels?
I just have doubts it would have a reputation and following if it was designed to sound and test this bad. Just thinking out loud.
 
at 3ma per section and 115v across it, that shouldn't abuse the tubes very much at all. Datasheet conditions at 100 volts call for 15 millamps of current for a single triode section. It also allows for a total 2 watts dissipation across both triodes, you're at 0.8 watts total.

its not impossible its just under-biased honestly. You could make that cathode resistor smaller and wind the current up some, looks like that would improve the linearity somewhat.

as for why it measures badly vs how people think it sounds, consider some folks just have tin ears or are allergic to measurements. This is about as simple as it gets for a gain stage, other than being parallel sections. I suppose if the sections were just grossly mismatched it may do funny things but there isn't too much here that would produce HF roll-off, other than the 100k input resistor and an excessively low impedance on the pre-out end of things. This won't be nearly as fussy about load as say a 12ax7, but its still only so capable.
 
I was testing unloaded (without amp connected) and could get past 5v peak without clipping on the sine wave. Then I ran through frequencies at 1v with the same result
 
well, that rules out loading and clipping I think. Especially if the scope was on x10, thats a very light load.
 
1x. I forgot that probes may be loading it and I’m possibly not seeing real results. Let me try with low capacitance probes tonight. I’ll post some squares too.
 
x10 low capacitance probes did not make an appreciable difference.
I guess there's no winning with the input pot and triode stage. The roll-off will be variable but will always be there. All previous scope tests were at volume full up.

Today I experimented with pot positions that will likely be used most of the time. The only line input source I ever use is my trusted Teac DAC which has its own volume control. I assume the volume is adjusted digitally at the DAC chip level and not in the analog output section of the device but I don't know for sure. Having variable DAC output helps with the preamp volume control setting. Plus there's a digital volume control within the interface, ITunes, Colibri, etc, The latter adds the convenience of being able to adjust the volume with the keyboard.

Here are 1KHz and 10KHz squares with volume at full max. The scope display is at 1v/div set at 0.1v/div at x10 probe. I know 10KHz is not expected to be looking as good as the input square but the one I got is awful.

1.jpeg

2.jpeg

While keeping the same output amplitude, the square display is better between 9 - 11 o'clock positions but without an input resistor bypass cap, it still shows a visible roll-off above around 10KHz when I sweep through the audio range with the sine wave.
47pF bypass cap makes response flat to around 30KHz at 11 o'clock but shows some HF boost at the 9 o'clock position, which may be desirable at lower listening levels. Better than the roll-off without the bypass cap I think.

Jumping input resistors doubles the gain and will probably add noise to the input. I think bypassing input resistors is less invasive if I decide to settle on that modification.

gadget73 said:
its not impossible its just under-biased honestly. You could make that cathode resistor smaller and wind the current up some, looks like that would improve the linearity somewhat.
Click to expand...

Do you still think that would be a reasonable approach? Maybe this preamp is known to be hard on tubes due to its high heater voltages? Since it's not regulated, with higher line voltages some people report at their outlets, heaters can get really hot. Mine was showing 6.67v at 122vac input. I've seen my wall go as high as 125.5vac. At some point some years ago I registered 129v but I don't think the meter I used then was to be trusted.

While all this is being percolated, I'll draw the phono stage schematic next. Then, if I will really feel like challenging myself, regulated HV power supply.
 
Also, any opinions on what was the reason to use such odd resistor values?
Just to make serviceability difficult for DIYers? I understand the OCD with 1% for everything but odd values...?
Plate resistors are bigger than 1/4" in diameter. What would be the wattage? I'm sure it helps keep them quiet with their heat dissipation and PPM ratings
 
grindfix said:
Jumping input resistors doubles the gain and will probably add noise to the input.
Click to expand...
Just the opposite. Those resistors increase the noise contribution caused by the excess gain within the preamp. A divider at the output (or other means of reducing internal gain, such as removing the bypass cap) will reduce noise at the output of the preamp. Did you ever examine the response with the 100K input resistor jumpered?

Jack
 
Just ran a quick square wave display and it looks better at typical volume settings. Higher gain but more linear and not falling off with volume pot adjustment as much. Will run more tests tomorrow
 
Also thinking to put some resistance from pot wiper to ground. That should also help reduce roll-off at higher pot positions, right?
 
grindfix said:
Just ran a quick square wave display and it looks better at typical volume settings.
Click to expand...
Was this test run with the 100K jumpered?
grindfix said:
Also thinking to put some resistance from pot wiper to ground.
Click to expand...
Is this on a PCB, so the circuit is difficult to change? I don't understand why they designed this with two sections in parallel, rather than driving only the first section and using the second as a cathode follower. That would have reduced the input capacitance. Anyway, if you remove (jumper) the 100K, but response still isn't flat enough, you can add a small capacitor from the top of the pot to the wiper. That's how it's done in many commercial applications, especially if the gain tube has high Miller capacitance, such as a 12AX7. Eliminating the cathode bypass cap will reduce gain and also help flatten response.

Jack
 
You must log in or register to reply here.
Back
Top Bottom