Audio Electronic Supply AE-3 Tube Preamplifier Repair/Test

[Smoky]

Hey Folks, a while ago I purchased this AES AE-3 tube preamp to use with a pair of AES Six Pac mono tube power amplifiers.

The AE-3 preamp was offered assembled or in "kit" form. This unit looks like it was sold as a kit (no serial number).

I connected the preamp to a Nakamichi PA-7 solid state power amplifier and JBL speakers and one of the channels sounds really muddy and I can hear plenty of hum. The 100K volume pot seems to work ok but it's definitely not "clear" sounding.

My first question is about a dummy load for the preamp outputs when testing it. I read that if a preamp is feeding a tube power amplifier a 100K resistor with a 300pf capacitor in parallel is used and if it's feeding a ss power amplifier it's a 10K resistor with a 1000pf cap in parallel.

Does that sound about right?

[Smoky]

Here's a picture of the underside of the chassis. The solder joints look dry/dull. These preamps were sold around 1999:



The schematic looks very simple showing just one channel:



I have looked under some of the newer AES and Cary Audio preamplifiers and they use shielded cables instead of twisted pairs plus the chokes were known to couple themselves to the main transformer on top of the chassis too. It looks like a few things can be consolidated/moved around/shielded.

[Wallace Gasiewicz]

I think most of these amps are made for 600 ohm output. That is whatt the cathode follower output is for, So a 10 watt 600 ohm resistor should be a good load. Big resistor to avoid too much heating...
Hum is usually caused by inadequate filtering. Check all caps and make sure the caps on either side of the choke are in the correct places. There is a formula for this filter setup Cap---Choke---Cap but I do not remember it at all, also a formula for the Cap-----Resistor----Cap part of the filter,  if the caps are not in the right place it does not work very well.
If you have a scope you can look for the 60 Hz hum...
They are using the same value for all the caps...I think this is odd.

[jonpaul]

dry leaky 23 yrs old lytics,

Check cap C, ESR, leakage, recap.
Check every PSU V, ripple

j

[Smoky]

Check this out guys, there appears to be more to the story!

I began feeding the preamp 1kHz at 500mV on both the left and right CD inputs simultaneously:



I noticed that when I disconnected the right channel's input the oscilloscope was still reading a small signal on the right channel's output

So I fed the signal only to the left channel's input and put the scope probe only on the right channel's output:



It looks like the L&R signals are able to cross over to each others wires inside of the preamp!

[jonpaul]

suspect bad grounding. perhaps cooroded ground lug

Never heard of this old brands but in the Audio world,

AES = Audio Engineering Société ,since 130 years ,it is a registered trademark!

(:-:)

j

[Smoky]

I believe your suspicion is spot-on JonPaul

The chassis has a powder paint finish and the screws are painted black. It doesn't look like any of it was scraped away either.

The metal bottom cover is the same way too.

So it's like the bases of the terminal tie strips are floating above the chassis. You can see the power cord's ground wire connection at the first tie strip. I'm sure no good ground goes beyond that.

On top of that, when I touch the preamps chassis with my hand while it's running, the noise dies-down quite a bit.

This reminds me of a couple weeks ago when I installed fresh ground rods for a customer. Their original rods were corroded, totally gone with flaking rust at the connection. I showed them that there was no conductivity. At best, the rods are made of copper-clad steel. Most are just galvanized. Ten years of useful life, maybe?

[floobydust]

Sorry but I dislike these designer audiophool builds - they take good, trendy components and stuff them in a box with a terrible layout.
This thing's twisted pairs are a rat's nest and will pick up hum and noise inside the chassis. Then zip tie them together, including the output lines 
The preamp has high input impedance 1MEG and even tiny stray capacitance will couple enough signal either between inputs/channels for weird cross-talk, or to the PSU for hum. It will depend on the source impedance and the volume control setting. And the ground loop with audio ground going to mains ground making a mess with other gear.
Twisted pair is not shielded cable and it's grounded at both ends - so I think the basic build is doomed.
The choke has an air-gap and radiates some B-field with 120Hz and its input is under an inch from the RCA's and the rat's nest running near it.

If you wanted hi-fi out of this, first off the input/output cabling needs to to be properly shielded and spaced, away from the AC sections of the power supply. Using equal length coax to the input selector or a steel metal shield along the edge of the chassis would work.
May be look at the original's layout, I'm sure it was done better than the kit build.
Why is the one Nichicon VX plastic wrapper torn like that? Seems suspicious...

[TimFox]

I think most of these amps are made for 600 ohm output. That is whatt the cathode follower output is for, So a 10 watt 600 ohm resistor should be a good load. Big resistor to avoid too much heating...
Hum is usually caused by inadequate filtering. Check all caps and make sure the caps on either side of the choke are in the correct places. There is a formula for this filter setup Cap---Choke---Cap but I do not remember it at all, also a formula for the Cap-----Resistor----Cap part of the filter,  if the caps are not in the right place it does not work very well.
If you have a scope you can look for the 60 Hz hum...
They are using the same value for all the caps...I think this is odd.

The output cathode follower seems to be biased to 3 mA quiescent current, so it would be unhappy driving a 600 ohm load.
Normally, such preamplifiers drive a high-impedance amplifier input, perhaps 100 k or more, but certainly no less than 10 k.

[Smoky]

Here's a tidbit of the instructions from the assembly manual. It shows twisted pairs but looking at the schematic it has that "loop"symbol which would mean shielded?



Floobydust, that capacitor's sleeve does look melted. All of those components are coming out so I'll measure it.

Btw, isn't that BR-102 10A bridge rectifier a little overkill for the two tubes? Two 6SN7's in series is about 1.2A right?

Anywho, I have a restructuring plan brewing for this little preamp

Also, here's a quote I found on the internet regarding one of the tests for preamplifiers and the loads applied to their outputs:

Frequency Response at Unity Gain with IHF and Instrument Loading

    Purpose: Gives an indication of how flat and uniform the frequency response of the preamplifier is and how this response varies with output loading. This test is done by setting the volume control for unity gain with IHF loading. A standard IHF input signal level of 0.5V is applied.

    What it tells you:  The comparative response with IHF loading (10k ohms in parallel with 1000pF) and with the instrument loading (100k ohm in parallel with 300pF) gives an idea what to expect with different power-amplifier input impedances and interconnect capacitance. Some preamplifiers have relatively small film output coupling capacitors and if connected to a power amplifier with low input impedance, some loss of low-frequency response will occur. If a long run of high-capacitance interconnect from preamplifier output to power amplifier input is used, the high-frequency response with the IHF  load will be relevant.

[floobydust]

Yes I would say the original schematic shows shielded cable (symbol) and I did see an AE-3 with shielded cables. The twisted pair is a mistake IMHO. Unless you stuff them in a copper pipe (or piece of square conduit) for the run from back to front, across the choke. Prove it with the scope check, if the pair is going to work.
For coax, problem is you want the same (coax) matched capacitance for both channels but the one furthest away is always a longer run and has more capacitance, which causes an imaging problem.

For the hum problem, I would select and short an input, say AUX1 and then see what the scope ACmV shows. It should be very quiet, low noise floor. Sometimes mains RF, AM radio etc. can have a hay day inside and give some pretty gritty scope traces.
Another method is covering the end of a scope probe on x10 with electrician's tape and probing the chassis components at a distance. You move the probe around to see where the dirty, noisy E-field is, and along the twisted pairs. 1N5408's I find slow as molasses and make a lot of switching hash.

I looked at a Cary SLP-50B and it's the same twisted pairs right next to the choke, and a bad ground with the transformer center tap.
These boutique builds, I find if you don't have basics solid as far as shielding and grounding, just sound terrible despite the expensive parts used.

[jonpaul]

been in pro, high end audio since 1970, knew Sol Marantz, Sid Smith, Dick Sequerra, Sigfried Linkewitz

Very sorry, but reading all of this, I Think this poorly designed and poorly built  amp is a waste of time.to restore

Many modern peramps, HP amps will far outperform for perhaps $100.

For a classic tube/ valve unit I would seek a known and respected vintage brand.

Our Swiss friends at Weiss Engineering, and also at Apogee, have fine amplifier modules that can drive headphones and lines 600 balanced.

Just the ramblings of an old retired EE

Jon

[Smoky]

I spent a little time over the past few weeks piddling with the AE-3 preamp. The reason why I want to get the best out of it is because I bought some tube gear from a late friend a few years back. He had two of these SixPac amplifiers stashed in the closet that needed work. I rebuilt the SixPacs but I wanted a matching AES preamp to drive 'em. So there goes the story.



It took a while, but recently I dropped over $500 for tubes to get the SixPacs up and running



The tubes came from Apex Matching. They take factory-matched tubes and match them even further. All twelve of these EL34 tubes measure the same in current and transconductance.



Coming up are a bunch of pictures showing what I did to the AE-3 preamplifier hoping to make it a little quieter/tolerable

[Smoky]

She's quiet now!



...maybe, a little too quiet

[Smoky]

This reply is for those who may have bought one of these AE-3 preamplifiers new as a kit or used.

If you look at the way the power plug is wired in my preamp, you can see how the hot wire after the fuse goes straight to the transformer and not to the on/off switch:



Looking at the assembly instructions, that's the way they wanted you to wire it:



If you look at the schematic at the top of this thread, it's correct. It shows the power after the fuse going to the on/off switch.

So be aware.  The moment you stick the plug into your household receptacle, you may be energizing the transformer!

[Smoky]

I started by taking a look at the power transformer.

As you can see, the transformer was potted on an angle with it touching its outer shell. You can see how I pulled back one set of wires out of the wax and set them to one side. That gave me an idea. More on that later:



So out it came:





I then trimmed off its legs and prepped it for re-potting:



I prepped the aluminum shell and base by grinding the powder paint off of the edges to make sure a good ground contact is made:



Back into the cauldron the tranny went. But now, with the wires prepped to one side:

[Smoky]

Knowing now that the wires are coming in and out of the transformer on one side, why not try to push the entire power supply within the preamplifier's chassis to one side too?

So I fabbed up some brackets by tack-welding some .06" steel in hopes of making that happen:



I also thought I could relocate the selector switch from the front panel all of the way to the back of the chassis at the same time.

So I made a bracket for it too.

The transformer got an extra layer of 16oz. copper flashing and a ground wire for added shielding and a 1/16" thick piece of foam rubber padding for vibration control (maybe):

[Smoky]

Looking back at the original layout of this preamplifier, you can see a plug filling a hole near the red dot on the top of the chassis. This preamplifier must of had an option for another vacuum tube or something:



Since I plan to cram the power supply into this corner of the chassis, why not use it as a vent?

So I found this cool stainless steel,  heavy-gauge fiilter on Ebay for a $1.83, so I bought two:



Next, I installed the transformer, the bracket, and then the filter:



...and there you go

[Smoky]

Next was to build the power supplies. My #1 goal was good grounding. I ran SPC wire throughout the preamplifier to all ground points. I did not rely on the chassis even though I made sure every connection made to the chassis was ground to steel, lightly wiped with dielectric grease, and securely fastened with lock washers.

FloobyDust, you mentioned "hash" noise made by 1N5408 diodes. I looked around Mouser and Digikey for what I considered the best to use. I may be wrong but I bought Vishay 600V 4A  VS-4ECH06-M3/9AT diodes for the high voltage side. Vishay claims these diodes do not lose their speed as bad as others when temperature rises (this tube amp may get warm).  I used Vishay 100V 3.5A SBYV28-100-E3/54 diodes for the filament side:





I used turrets to fasten the transformer wires to make it easy to fix the power supplies if needed in the future.

[Smoky]

Next was to cram it all into that small space.

I added a small board to the main board to take two 47K resistors away from the crowded tube socket area:



Next was the filament board:



And, finally, the choke:

[Smoky]

Next is to install octal sockets that are "slightly" older than me



Now, some of the base wiring can begin. There will be a future wiring adjustment made later.

[Smoky]

Next, it was time to install the socket board. All of the parts were installed ahead of time. The leads are pre-bent and fall right into place:



On the back, you can see that all leads to the socket can flex when tubes are replaced:

[Smoky]

It was at this point that I could take some voltage readings.

All of the high voltage measurements were close to being spot on. I felt some voltages were dependent upon good and matching 6SN7 tubes which I believe the ones that I have aren't.

It was when I measured the filament voltage, it seemed low. 11.76VDC was exiting the L78S12CT (TO-3) 12V voltage regulator which gave each heater for each triode within the 6SN7 just 5.53V instead of 6.3V. I read that any measurement +/- .3V can cause noise or accelerated aging. Again

So I made a judgement call. I stepped up to a 15V regulator. An LM140K (TO-3) from 1978. It requires good heat-sinking to go above 1A. Each 6SN7 requires .600mA plus the diode on the front. So I guesstimate ~1.25A of current is being devoured. Just over 1A. Well, this preamp has the regulator mounted on top of the chassis with a large heatsink so I went down the 15V road:



I installed the LM140K along with the recommended transient suppression capacitors from their datasheet:





On the bench, the LM140K measured 14.79VDC meaning I may need to drop ~2VDC!



What do I do? Well, how about a dropping resistor

Can I Well, I did. I calculated a 1R5 at 2Watts. So I bought a 1R5 5Watt cement resistor and made a spacer/mount to go under the socket board. I layered it with brass and made a spring to keep it pushed up on it's hold down clamp in hopes of getting air around it better:



And here it is installed. It may look upside down but when the preamp is upright, the heat from the resistor will flow up towards the underside of the chassis, not the Phenolic board, well, cross your fingers that it will



And the result, and after three hours of idling, I measure 6.5VDC per heater. I measured 45C or ~110 degrees Fahrenheit on the Phenolic board with a Fluke IR gun. Plus, the heat of the tubes is rising in the wrong direction during the test. I'll keep an eye on it:

[Smoky]

The relocating of the selector switch is next.

I wired the new switchcraft RCA jacks first with equal lengths of wire:



This picture gives you an idea of what these six short lengths of wire are replacing:



The selector switch mount, the selector switch, the black Essentra shoulder bushings, and the RCA jacks were next. The switch is a 2-pole, 3-position C&K silver-plated contact type (A20305RNZQ). The ground wires are separated from the left and right channels at this point. Only one input can work at any one time anyway:





The 6mm CNC belt came next:







Followed by the new Alps 100K volume pot and board plus the pair of equal lengths of RG316 coax from the selector switch to the volume pot and then from the volume pot to pin 1 of the octal sockets:



It's getting there

[Smoky]

I'm using Mundorf MCap Supreme coupling capacitors:



The RG316 coax was soldered to the RCA jacks and shields installed afterwards:







The shields were placed over the .22uf capacitors near the tube sockets and the equal lengths of RG316 were attached to pin 6. The Mundorf caps were then dropped in next:





The brass cover was then tack-soldered in: