Blue Angel issue - the verdict. Amp is not the issue.

I spent the better part of the weekend with my Amp Tech - Great guy by the way - Dean Zink in Sylvan Lake, MI.

Turns out - it wasn't the amp at all.

First I isolated the issue to the EL84 side of the amp. Then Dean went to work.

Parasitic Oscillation - induced by not having enough proper load on the amp. I use an attenuator to push the amp into overdrive... and it turns out you can't reproduce the sound just playing into a cabinet without my attenuator.

So, I took my attenuator over today and we finally reproduced the issue... and managed to also reproduce it on his scope using a signal generator.

It appears that the Blue Angel has a few things it doesn't like by design. One of them is variable signals while attached to a dummy load. Just using a signal generator you wont see the issue - you have to rapidly tweak the volume knob to induce a signal spike that starts an oscillation... Or play a guitar... You see typically, an amp tech will check the output for stability with a nice stable signal generator... problem is, parasitic oscillations like this don;t show up without a varying signal to induce the oscillation.

Its actually quite fascinating - I spent three hours in his shop as he picked the entire amp apart - modifying feedback resistance, adding output load, adding and removing snubber caps - all trying to filter the oscillation when my attenuator was connected. At the end of the day - you just can't get rid of it all.

What we ended up doing was a fixing a couple design flaws in the Blue angel design.

The EL84 tubes are run with incorrect screen resistor values for some reason - he thinks mesa was trying to drive them harder to force them into saturation... but the problem is that it appears that they are just running way too hot because of this. So he modified the resistor values to get them back to a reasonable level.

Second - he told me that he generally likes to keep the feedback at about 50% - and the Blue Angel was is set up around 75%... so he modified this as well. He said it may be that Mesa was trying to improve clean headroom, but as you increase the feedback level too high it can induce issues in the bass response and make the amp sound sterile.. ( I hope I'm remembering this correctly).

Finally - he modified my attenuator. My stock attenuator is an inductive load (a kit I got second hand in a previous amp purchase). He found that when he added a load resistor (actually two) that it helped stabilize the oscillation a bit... so he added that to the attenuator..

All in all - the issue is still there - although to much less extent... And ONLY if I use the attenuator and push the volume up past 50%.

His parting words were that unless I want to really get deep into the amp and change the output transformer to rework the power section, there isn't much that can be done. He said its a great amp for regular guitarists that just want to play through a cab - and arent loony like me, trying to push the amp to 10 all the time!! He said that Mesa's tend to be a bit finicky and don't like what I'm trying to do. I give him credit for going to such extents to diagnose and try to eliminate the problem - he was very very fair in his pricing and talked me through every step of his process.

A side note for the tech geeks:
The Blue angel runs both 6V6 (2) and EL84(4) tubes. It does NOT, however, use separate taps on the transformer to run these tubes. Basically, Mesa uses a transformer that is a lot more powerful than would normally be required for a 35 Watt amp. He said they basically "cheated" and over spec'd the amp and run both sets of tubes from the same tap... Since the transformer is enough for about 50 watts, it can handle the load... but it is probably creating some of these problems.

There is also another cheat he found in his searching - there is a cap with a wire wrapped around it... I thought it was a choke and he told me that the cap most likely is getting (or did he say sending?) interference to one of the traces on the board - so the designer wrapped the wire around the cap to form a shield from this transmission.. I thought this was very interesting.


Edit to add: By the way - if you noticed my add in the classifieds for this amp yesterday evening - I apologize for that - I was so frustrated because at that time we couldn't figure it out that I posted the add in a bit a huff. The add has obviously been deleted... Not sure whether I'm keeping the amp or not, but I'm at least satisfied that its not the amp's fault.

You didn't mention the attenuator in the other thread. If you had, I would have said one thing straight away...

The Blue Angel does not like attenuators!

To be fair, the sound you were getting didn't sound quite the same as mine, but maybe I should have asked.

If you really need to use one, try a purely resistive type (eg Airbrake) - the only one I tried that sounded OK with mine was an own-built pure resistance load. This does figure, from what your tech found out. If I remember rightly mine particularly hated the Marshall Powerbrake, which is probably the most inductive of all attenuators.

You may also want to try some different tubes - mine really worked a lot better with a 12AY7 in V2 (which removed the annoying hiss and some of that over-mid-focused tone, although it does reduce the gain a bit too) and 12AT7s in V5 (PI) and V3 (reverb), which deepened the tone, cleaned it up a bit more, and made the reverb less crashy.

I think it's a really good amp, but perhaps not quite the way it comes stock.

Yup - I wish I would have posted that as well... In my (previous) naivete I didn't understand the different types of loads, i.e. resistive v.s. impedance... I got a thorough education yesterday!

So - that being said - a Weber MASS style should be ideal for this situation, as it uses a real speaker drive as the load, correct?

My understanding now is that since a speaker only provides its rated load (4, 8, 16 ohms, etc) at a very narrow frequency bandwidth, it also may have huge loads at much higher frequencies (i.e. 20Khz and above)... We noticed this when we first scoped the amp with my attenuator - supersonic frequencies made it lose its mind. Not so when it was connected to the cab or a resistive load. So a speaker is essentially a varying load based on input frequency.

The Weber does use a speaker voice coil, but only as part of the load - I don't know how much, proportionately. I haven't tried the full-size Mass, but I had a Minimass once which was by far the worst-sounding attenuator I've tried. (And was poorly-built, too.)

What's odd is that the Blue Angel doesn't mind a real speaker, which is highly inductive, but does mind an inductive attenuator. I have no idea why! But the Airbrake or another pure-resistive attenuator should work OK.

Wow! Very informative post! I wonder how many other amps out there gets the parasitic oscillation. That would be a good band name, "Parastic Oscillation" could be a mouthful though.

After reading that post, makes me wonder about amp designs. Your tech appears to be very experience, educated, and savvy. Most of all, seems very honest. That's a good find in itself there.