Ashdown BTA-400 Bass Amp Severe Clipping

[nevetslab]

In the heavyweight class of Bass Amps, Ashdown’s 400W BTA-400 ranks among them, with it’s eight KT-88’s and reasonably large Power & Output XFMR’s. The one in our inventory came over to the shop due to output level dropping to nearly zero and severely distorting.

While playing thru the amp, it’s output spiked a couple times when I hammered aggressively on the bass, enough to cause me to move it to the bench, attach it to my load bank & scope to watch, and not risk taking out any speakers. After 10 minutes playing, I heard something click, and now had severely distorted and clamped output level, while it was still drawing substantial AC Mains current/power (5-7A) under drive conditions.

Pulled it from the case, set it upside down with the Xfmrs on a pair of 2 x 4’s, and began inspection. Nothing looked out of the ordinary, beautiful construction. The Power Tube PCB is in the form of a horseshoe, with a driver PCB Assy parked in the middle above it, Power Supply PCB on one side, and bias/screen resistor PCB on the other. Preamp section in a separate bolt-on chassis that can be detached, while remaining tethered by the wiring. Bias system has an LED Bar Graph and a dip switch, allowing each power tube, with it’s own bias control, to be set individually, not requiring a set of 8 matched tubes..

I powered it back up, 4 ohm load attached, no signal applied. At idle, it draws around 3.1A, and all the power tubes were nominal (around 29mA ea). Driver PCB uses a pair of ECC82’s and an ECC802, configured as a gain stage, phase splitter, AC-coupled gain stages and cathode follower stages to drive the output tubes.

PAGE N DRIVER MODULE.pdf
POWER AMP OUTPUT MK2) PAGE L.pdf

When I applied signal, the output was still clamped low, severely clipped, and would draw power if pushed.

Pulled the power tubes,, lowered the AC Mains to drop the supply voltages back to around 500V(driver stage level) and began probing the driver stages. I had sensible signal thru the power amp input stage & phase splitter, but while having signal on both grids and cathodes of the final gain stage following the phase splitter, I no longer had output on one of the plates.

Swapping tubes (2nd ECC82) yielded no improvement, and began getting intermittent output and oscillation, before the one side would suddenly stop conducting. Swapped driver tube, no change.. Left the cathode follower tube out all together, and after setting the new 12AU7’s aside, as both seemed to make matters worse, stayed with the original tubes. At one point, when I lowered the Variac a bit, I had output looking normal, and as I raised it up, the ‘bad’ side began breaking up and then stopped conducting all together. So, at least I had found WHERE the problem was, though not WHAT the problem was.

I re-soldered the heater & plate terminals of that driver tube, which looked totally fine, no change, so it wasn’t an abrupt loss of heater current thru the one side. Plate resistors were fine, common cathode resistor, and the other side in that voltage gain stage was fine, so noting wrong there. I didn’t see any breakdown of the coupling cap to the ‘bad’ side. Replacing it would be no simple matter, unless I just snapped the radial box cap off. Getting that driver board removed from the assembly is major surgery.

Almost by accident, when I had changed the ECC82 with a 12BH7 (used in place of the ECC802), I had bumped the cathode follower tube, and found I could cause the signal from the previous stage to cut in and out. Close inspection of that tube socket soldering revealed nothing…totally healthy looking. Nevertheless, I de-soldered and re-soldered that tube socket, and tried again.

That cured it! Output now solid as a rock. Restored the ECC82 to the middle tube, still fine. I had installed a 47pF on that stage plate-to-plate to counter the oscillation I had seen previously, and left that.

Restored the output stage, and all solid as a rock. Still mystified why I was seeing high current drawn before when the one ECC82 gain stage would drop to zero, and the output wasn’t just producing half-wave output, both sides of the output wave was only around 1V, severely clipped.. Burned it in for solid hour using burst pink noise, and later playing thru it, totally healthy.

Solder joints. Totally healthy looking, and yet behaving intermittently. AND, in the following stage. Even with that cathode follower stage tube removed, I still was having the gain stage abruptly stop conducting. Weird one!

[Tom Phillips]

Nevetslab,
Thanks for taking the time to write these detailed posts. I enjoy reading them.
Cheers,
Tom

[nevetslab]

Output into speakers now horribly distorted

The Ashdown BTA-400 I had restored recently just came back, now unable to drive loudspeakers without making severe 'farting' sounds, not unlike making such noises with your lips as we all did as kids, but does a remarkable resemblance to that playing thru it. I didn't hear this initially Thursday when it was wheeled back into my shop, and it being needed for the American Music Awards show that night. I pulled it apart to check on what I had done previously, and all seemed ok. Never got it to 'hickup' by pounding on it, that being one of the symptoms reported. I didn't like the sound of the bottom end, but I also didn't have a speaker in the shop that could handle the power. I wheeled it back to the Guitar Dept and plugged it into an 810 bass cabinet, plugged in the bass and played it hard. Useless rude body sounds coming out of the speakers instead of solid round bass tones.

Dragged it back to the shop, dis-assembled it again, it sitting upside down, exposing the driver tube PCB, power tube horseshoe board and power supply board.

The point of break-up appears to happen when the drive signal from the cathode followers begin asymmetrically clipping. The stage ahead of that will swing about twice the amplitude, clips symmetrically. It's DC-coupled to the ECC802 cathode follower stages (see schematics attached in 1st thread). I don't have the amp on hand at present...I had to put it back together and send it back out as a stage prop, so it will be next week before I can post scope photos.

I still don't have any ECC802 dual triodes on hand, which is the tube called out for that cathode follower stage. Presently using a 12BH7 there, tired both I had on hand, tried 12AU7's. Swapped out the 12AU7 stages ahead of the clipping driver stage, with no change. I did find the 47pF cap I had previously installed across the differential plate drive circuits to the final cathode-follower driver stages was causing the output to oscillate when it lost symmetrical voltage rive, and that got rid of the oscillation bursts, but hasn't stopped the drive problem

I've actually never seen a BTA 400 that works correctly, only the one we have that didn't, and did at least cure the initial ailment. I have solid output at 200W/8 ohm from 400Hz down to 30Hz into a resistive load bank, so I don't see a problem with the output transformer. What I've tried so far is removing the power tubes, after first verifying all were biased up ok, Then, with the power tubes removed, disconnected the bias voltage feeding all 8 bias adjust networks, four per driver stage, just to remove any DC leakage issues, while monitoring the cathodes of the ECC802 drive stage. Then, finding the actual circuit on the PCB not being identical to the schematic, there isn't a series 15k resistor ahead of each bias network, so no simple way to remove the eight bias networks off the cathode's coupling caps. I've increased the cathode resistor to 100k, lowered it to 15k, neither having any corrective change on the clipping level. No faulty component values found in the bias networks or in the cathode follower stage. I tried a brand new set of KT-88's, biased them all up, and still have the severe clipping condition. Tried different 12AU7's, different tubes in he cathode follower stage

It looks like the speaker output distorts and loses control of the output when the driver stage clips, and the output then spikes with the horribly distorted sounds. It's now happening at low level as well, into a speaker. So, something has changed since I've been trying to find what is at fault. I don't see a lot of difference in the tube specs between an ECC82 and ECC802. The 802 has a little more current drive, and it looks like the 12BH7 being used has about the same, while having a higher plate voltage rating an slightly lower plate resistance. Doesn't appear to be tubes itself as the problem.

It behaves like the back emf from the loudspeakers, when driven out of the magnet gap is causing a current limiter circuit to bark, a not-uncommon problem on many solid state amps having untamed current limiter circuits. Only, of course, no such circuit exists in this tube amp. Just behaves like it. I do see an oscillation spike when it loses linear cathode drive to the output stage....monitoring both driver and output on the scope.

I should have the amp back in the shop this week for more mental abuse. Nothing simple in regards to pulling the driver stage apart. It is hard-wired from that board down to the horse-shoe power tube board below. Only way to separate the 2nd stage plate-coupled DC drive to the grids of the cathode-follower stage is cutting traces on both sides of that driver board. Ashdown still owns the amp, it and others in our inventory...parked here for exposure with our high-end clients.

[nevetslab]

Output waveforms & power amp images

Having recovered the BTA-400 from the American Music Awards show this afternoon, I pulled the amp apart again, and got the following photos of the chassis, set upside down sitting on 2 x 4 blocks under the transformers. From the rear view, the PCB assy on the right is the power supply board, the board on the far left is the Bias Meter and Screen Resistor board, the upper board in the middle is the input, phase-inverter, buffer, cathode follower stages and all eight of the bias networks, one per power tube. Below that is a horse-shoe-shaped board which is the power tube PCB. All hard-wired together. The preamp assembly is the enclosed chassis forward to all these boards, which I haven’t removed for the first set of photos. The last two power tubes are just forward of the middle PCB assy.



After taking these images, I set up the amp, driven by 50Hz 1/3 Octave Pink Noise (Bruel & Kjaer 1027 Sine-Random Gen + B & K 1617 1/3 Octave Filter Set), with the amp connected to an Ashdown 810 speaker cabinet via 20 ft of 12AWG cable. Scope probe connected to the positive-going cathode follower driver tube and output from the amp connected to my Amber 3501a Audio Analzyer, output fed to one channel of a Tektronix 7633 Storage Scope, with both waveforms displayed and captured in storage. The leading waveform is the Amp output waveform, the higher-magnitude waveform having oscillation bursts is the cathode follower stage.



During the waveform capture series, I increased the output level to where the amp was finally clipping, as seen in the 3rd and 4th scope images. Oscillation isn't passing thru the power tubes/Output Xfmr, but sure present at the CF stage. Last scope image was at a lower level, but still getting this random severe clipping.

After I had captured a number of output waveforms, I shut the amp down, removed the eight KT-88 power tubes, and added an additional scope probe so I could look at the input grid and cathode output of the cathode-follower stage, using a 12BH7 in place of an ECC802, not yet having one on hand. I powered back up, now just looking at the driver tube input and output, without the power tubes installed. I changed the signal source to 400Hz Sine. Both X10 probes are set at 50V/Div (CRT Readout on the CF waveform doesn't always read correctly). I can't help thinking this clipping on the negative half of the CF stage is the cause of the inability to get solid output from the amp. Both CF stages look like this.



I discovered this time the driver stage was fully engulfed in oscillation, which I wasn’t seeing last week. I swapped 12BH7 tubes, no change. Swapped the PI buffer tube (middle ECC82 tube), and the oscillation went away. That’s the clean waveform like I was seeing last week. I quickly put the power tubes back in to see if the problem was still there under drive. It was. I ran out of time, didn’t get any additional scope photos with that middle tube replaced. I’ll look at that in the morning, but, the amp still craps out under drive.

Interesting note about the 50Hz 1/3 Oct Pink Noise test signal. The waveform appears as a random-amplitude LF sine wave, where the magnitude excursion differences can be as much as 10dB, I think. I find it very handy in checking LF subwoofer cabinets, power amps driving them up to over-excursion to see how the current limiter circuit behaves with the back-EMF energy feeding back into the amp's output stage (on solid state amps), and it does an adequate job showing up this fault condition without having the bass guitar strapped on Although the bass produces a much more dramatic drive signal, showing the failure of this amp’s ability to drive speakers in its’ present sick state.

I'll resume the efforts tomorrow, but that's what I'm up against on this beast.

[MarkusBass]

Have you got the rest of the schematics (especially the power supply)? The PI stage has voltage followers on the output. They are supplied with 450V. Not every tube can work correctly in such conditions if the heather power supply is not elevated. Can you check what voltage is on cathodes of ECC802? They could use a MOSFET as a voltage follower.

Mark

[nevetslab]

I have the power supply schematic, but nothing else on this mode. I've attached the power supply schematic, along with a couple of the PCB component guides. That of the bias/screen resistor board is not the same as in this chassis, nor is the PCB component guide for the driver board, which I haven't attached.

2nd stage ECC82

Pin 1 205V Pin 6 224V
Pin 3 12.1V Pin 8 12.1V

Cathode Follower ECC802 (12BH7 installed)

Pin 1 501V Pin 6 501V
Pin 3 230V Pin 8 238V

The Heater's are not elevated. So, the cathode follower stage has 271V & 263V from plate to cathode, with grid bias levels of -25V & -11V for the pair. AC Mains is 120VAC, with the KT-88's all installed.

I'm just ordering some ECC802's as well as ECC99's today. When I first saw this amp a year ago, it had the 12BH7 installed in place of the ECC802.

When I first serviced it in April 2014, there was an ECC99 installed for the Cathode Follower stage. I had seen the clipped waveform with that tube as well back then. I had replaced a cathode resistor on one of the two CF drivers, cured a bad solder joint, and had it working again. No idea where that ECC99 went, as it was still installed when I finished with the amp.

[MarkusBass]

I don't know the value for ECC802 but ECC83 has maximum allowed voltage between heater and cathode equal to 180V. You have much more in the amp. I wonder if this is important. The problem could be easily solved with MOSFET voltage follower (e.g. IRF820).

[nevetslab]

I don't know the value for ECC802 but ECC83 has maximum allowed voltage between heater and cathode equal to 180V. You have much more in the amp. I wonder if this is important. The problem could be easily solved with MOSFET voltage follower (e.g. IRF820).

The most common spec is listed as Uf/k, which I'd take to be the Heater-Cathode Voltage Max rating, DC. GE lists that for the 12AX7 as 100V, and DC + Peak of 200V, pos or neg. All the other spec's I've seen on the ECC802 is Uf/k Max of 100V, as is that for the 12AU7 & 12BH7. The ECC99's Uf/k Max rating is 200V. Like you, I too don't know if this is an issue.

I learned this amp was out with a client for 6 months last year, and had been taken to someone else for service, being returned with the 12BH7 tube. That explains what happened to the ECC99 that was installed when I first looked at it in April 2014.

The MosFET Source Follower idea is intriguing. There is space above the driver board, though not sure on the near-by fasteners that would be needed to mount a small board. I can just hear the sound of a nut and washer dropping into the depths below by tempting fate. building a circuit that's safe and sane, mounting to a 9-pin header to plug in from the tube side would be possible, though potentially dangerous if not properly done.

I'm getting to where a call or email to Ashdown on this is in order, as it still is their property. Love to what I'm overlooking.

I did just de-solder and re-solder the two ECC82 stages, having found the solder joints that looked totally fine on the ECC802 last month cured the signal drop-out problem on the preceeding stage. Made no difference though.

Seeing as the Driver board's supply voltage is running higher, dropping that series resistor to lower it might be worth a try.

[Enzo]

Getting to the point? Why not call them the first moment you feel the need of a schematic?

[MarkusBass]

Building a circuit that's safe and sane, mounting to a 9-pin header to plug in from the tube side would be possible, though potentially dangerous if not properly done.

Just for test, I would solder MOSFET to 9-pin header. This would allow for checking whether the voltage between the heater and cathode is an issue. And Ashdown could adopt your solution in their amps .

Mark

[g1]

I'm kind of surprised that they are running cathodes at 230V and roughly same H-K voltage.
Could this be part of a fault?
I know these kind of ratings are often exceeded, but this seems a bit much. As far as I know, ECC802 H-K rating is 100V.

[nevetslab]

I've received a reply from Ashdown, after sending a brief note of the severe output behavior driving any loudspeaker load. It sounds to them like the output transformer is flashing over above a certain drive level, either initially caused by client failure to have speaker connected, or a faulty O/T itself. Early in production they had issues with output tubes, and have settled with J/J KT-88 power tubes, and also had some defective O/T's. Once those had flashed over, it would continue to do so under drive above a certain level, resulting in a loud bang thru the speakers. Describes what I'm hearing. There were also mod's to the driver board in dealing with the problems I had described. The C/F driver tube they current use is an ECC99, which is what was installed in the amp when I had serviced it in April 2014. The amp had been out with a band for 6 months, and had been taken in for service somewhere during that time, returning with a 12BH7 in its' place.

I haven't yet discussed the high Heater-Cathode voltage being run. I did send them photos of the chassis, the S/N & scope photos after I got their first reply, so its' sounding like there are modifications required on the driver board, and most likely replacing the output transformer, after further conversation takes place. I haven't heard back since my reply to their response.

[nevetslab]

Parts are now on order to add circuit mod's that were never done on the amp we received. I won't know if those changes, together with the ECC99 cathode folllower installed cures the loud barking in the speakers under normal playing condition. They also disconnect the feedback loop now...still in place on this one. I should know by early next week if the output transformer is faulty. Sounds/looks fine driving a dummy load all the way to clipping.