The HODS project

[romberg]

So, I finished the enough connections to the board to scope out the skyline tone stack. And it seemed to work! Almost...

Things appeared to work ok, but only with the mid-boost switch engaged. When it was not engaged the PAB (middle position of rock/jazz switch) fell on its face and did nothing. Also, the mid-boost seemed really really strong. Like it was boosting the signal by a factor of five or ten. I considered that perhaps this was normal. Then dismissed that idea after reading some more on the site. No one else mentioned symptoms like this.

When the PAB was wonky and not working the signal sorta looked like it might be floating. And I knew it was related somehow to the mid-boost switch. So, I started connecting some things to the 4.7m ground reference resistor and fixed it! Scratching my head I examined my layout and my wiring matched it: But things only worked with the addition of my new magic jumper. The layout I used for the build looked like so:

mid-bad.png

Here the yellow wire on the left goes to the treble pot and the one on the right connects to the coupling cap of V1a. Folks who are sharper than I was will catch the error now. I however remained confused as the wiring matched my layout and I had checked that layout against the schematic a half dozen times! So, I went down a rabbit hole wondering if the schematic was wrong. But it seemed to be the same as 183 (minus the 22m pop resistors I left off). I was thinking of adding them when I spotted it. My layout did not
match the schematic!

The output of the "treble" coupling cap only connected to the pot when the mid-boost was engaged. Which explained everything! How could I have missed this a half dozen times? Checking on another computer I used to draw the layout I found this:

mid-ok.png

Dang. That one small little connection somehow got dropped between the time I checked (and double checked) things and the time I printed a large layout for the build. These things sure are complex and lots of little devils living in the details.

Mike

[romberg]

Finished up connecting the last tube socket. Still some work left to do. But I felt it was time for this pile of parts to become a guitar amp.

I connected it up to my dummy load, variac and current limiter. I had already verified the rectifier output. So, now was the time to check the main power rail. This is the first time I've done a start up with a variac. And it is way less stressful than just using a current limiter and flipping a switch (while crossing your fingers and holding your breath .

I connected DMMs to each side of the power rail and very slowly increased the AC voltage into the amp while watching the voltage slowly increase on the rail. Brought it up to full mains voltage and everything looked good. No tubes in so the rail was all the same potential. But no shorts. So far so good. Time to install the tubes.

Set the bias voltage to full maximum negative voltage on pin 5 of each power tube then installed all the tubes. Then I ramped up the variac again from 0 to 120 with the standby off (rectifier output disconnected from rail). I wanted to check that the filament wiring was still good. All tubes light up. Time for the real deal.

I cranked the variac back down to zero and engaged the standby. Then slowly ramped back to 120 while watching my two meters on each end of the rail and the third connected to my bias check points. It came up smoothly and everything looked good. about 450 volts on the plates:

startup.jpg

Most of the signal chain I had already traced with my function generator and scope. Now it was time for a full end to end test. Set the bias to 60% max dissipation and then connected my signal generator to the input. Here is a sin wave entering the normal input and the output into my dummy load. It is passing through all four gain stages and both tone stacks. It lives!

in-out.jpg

Yes the very next thing I did was grab a guitar and plug it in. I just ran the line out from my dummy load into a powered studio monitor.

line-out.jpg

It was kinda late and I did not wanna drag out a cabinet and annoy the neighbors. Even with this setup I can tell that this thing sounds pretty different from every other amp I've built. And I can tell it is going to take a while to figure out all the ways to dial in different tones. Just playing around briefly with it I can make either one of the channels overdrive the PI. And the OD channel can be run with or without PI distortion depending on how the masters and clean channel are set. Lotsa options even before one even tweaks the tone stacks.

Still need to sort out the voltage divider and install the FET board. And the negative feedback is still disconnected. The OD and PI trim pots are still just set at 50%. They may need some attention. The phase of the OT and PI input will require checking and the negative feedback connected. But for now my pile of parts is a guitar amp!

Mike

[romberg]

Working on the remains of the day. Checked the input to the PI and the speaker outputs. They were in phase.

feedback1.png

So, I swapped the wires from the PI coupling caps to my PPIMV and all looked better:

feedback2.png

Now they are out of phase, I hooked up the feedback connection to the circuit board. The presence control now works and the amp has negative feedback.

Next up was the FET board. I messed around with some resistors and alligator clips and determined that a 150k/10k divider would give me close to 20V. Which it did when there was no board (and other components drawing current). Once installed into the board and temporarily wired up, I was only getting about 10V instead of my 20. Foiled again! Some further resistor swaps found me at 20.5 volts with a 150k/18k combination. The board is clipped in and running (in what I'm sure is a non OSHA approved setup):

fet.jpg

Since all the current drawing components were now doing there thing, I checked some voltages. I was getting 175/193 on the plates of V1, 177/197 on the plates of V2 and 295/298 on V3. My power supply rail was 468-466-443-332-324. The voltage on my first two tubes looked a little on the low side. So I checked the schematic for 183 which while not a HRM is a high plate like my amp.

https://ampgarage.com/forum/download/file.php?id=16829

My voltages looked to be in the ballpark. Thanks for the suggestion to put the 22k, 2.2k resistors into my power rail! That seems to have worked out and saved me some more debugging trouble!

Here 183 is running a bit more voltage on the plates of V1 and V2 (192/198, 195/201). It's power rail is at 465-464-441-324-316. So, I had a bit more voltage on my last two nodes yet my plates were running at a lower voltage. Scratching my head I double checked the plate and cathode resistor values. They were right and matched what 183 has (220k/3.3k and 150k/2.2k).

I did some more reading here on TAG and found some suggestions to try different tubes. So, I went to my pile and found a pair of reflector 12ax7s that bumped up the voltage all on their own! I never considered a mere tube swap could make a 20-30v difference on the plates. I wonder if this is what folks hear when they compare different preamp tubes? It was a new thing to learn for me. After the tube swap my voltages look pretty close to 183 to me:

Code: Select all

     183  HODS
V1a  192  200
v1b  198  194
v2a  195  208
v2b  201  197
v3a  304  303
v3b  309  293

A    465  465
B    464  464
C    441  442
D    324  335
E    316  327  
Fet:      20.4/10.5

Next thing to do is mount and wire in the FET board. Then I need to figure out how to look into adjusting these trimmers. I just set them at 50% as a place to start. No doubt some tweaking will be worthwhile.

Mike

[martin manning]

Checked the input to the PI and the speaker outputs. They were in phase.
So, I swapped the wires from the PI coupling caps to my PPIMV and all looked better:
Now they are out of phase, I hooked up the feedback connection to the circuit board. The presence control now works and the amp has negative feedback.

The speaker output should be in phase with the PI input for negative feedback since it is fed into the opposite side of the PI.

Next up was the FET board... Some further resistor swaps found me at 20.5 volts with a 150k/18k combination.

A 150k/18k divider will probably work fine, but you don’t mention the FET bias. Did you check that? There is now an updated FET bias document, BTW.

[romberg]

Checked the input to the PI and the speaker outputs. They were in phase.
So, I swapped the wires from the PI coupling caps to my PPIMV and all looked better:
Now they are out of phase, I hooked up the feedback connection to the circuit board. The presence control now works and the amp has negative feedback.

The speaker output should be in phase with the PI input for negative feedback since it is fed into the opposite side of the PI.

Thanks Martin! I'm going go go and flip those PI output wires around again. Then re-read Merlin's stuff on feedback. Yiou saved me some confusion and head scratching when I would have fired it up and it screamed like hell. I appreciate the correction!

Next up was the FET board... Some further resistor swaps found me at 20.5 volts with a 150k/18k combination.

A 150k/18k divider will probably work fine, but you don’t mention the FET bias. Did you check that? There is now an updated FET bias document, BTW.

I measured the FET related voltages and jotted them down last in my little voltage comparrision chart. 183 did not seem to have any voltages for these.

I'm getting 20.4 volts at the junction of the 150k and what is marked Rd in your drawing. And I'm getting 10.5 volts at the junction of Rd and the drain. I had set Rs earlier in this thread using your 9v battery method.

I remember a bit of a discussion you were involved in regarding HAD perhaps biasing the FET differently. I'll check into your updated document. Thanks for the heads up!

Mike

[martin manning]

Ok, missed those voltages down at the bottom. You should have a nice clean boost with that bias point.

[romberg]

The FET is working and is a clean boost. In all the video/sound clips I've seen of real Dumbles the FET always was clean sounding. I'll admit that I've only seen three examples. So, small sample size . My plan is to dial mine in so that it boosts the lower output fender single coils up to match my humbucker equipped guitars. There does seem to be enough gain in the FET to cause some overdrive in the clean channel.

I fit the FET board in right next to the input jacks. Tight fit:

fet.jpg

I included effect loop jacks mostly because it was the number one comment I got when showing folks images of my panel design. I don't really use effects much and was planning on just not connecting them. But then I had a change of heart and wired them up. So, now all of the soldering is done. Here are some pictures of the finished chassis inside and out:

guts.jpg

front.jpg

back.jpg

top.jpg

Next I need to dial in the trim pots. I found some advice for setting the trimmer on the PI. I guess the FET trimmer is to taste. Maybe the OD trimmer is as well? I'll have to search around for advise on it. I'm guessing you set it to adjust the range of the OD gain pot with say the normal volume on around 5? Not sure.

Mike

[romberg]

Next up is setting the trimmers. In the following setup I have the rock and HRM switches engaged. All tone pots set to 5 as well as the presence. I'm injecting a 100mVrms signal into the normal channel at 1kHz.

First I found the max clean power I could get out of it into a 16 ohm resistive dummy load.

fullpower.png

Channel 1 is the PI input and channel 2 is the speaker out. The PPIMV is on 10, drive1 is on 5 and level1 is on 6.2. 28.3Vrms into 16ohms is 50 watts. I'm pretty close to that. The exact point at which the output starts to clip is subjective. Note that the pre-amp is still clean and there is more room to overdrive the power tubes and PI (more on that later).

Once I confirmed I'm getting 50watts max, I can set the PI balance pot. From what I understand from the sticky thread on this subject, I need to set this while the amp is running at half power using a sin wave of either 60 or 100Hz. I chose 80Hz as a nice in between value and lowered level1 until I had 20Vrms on the output (25 watts). Then I hooked a DMM and scope probes to my bias test points (cathode side of each 1 ohm bias test resistor).

I guess the idea is to get a zero reading on the DMM (set to measure mV DC). This was a little tricky as the meter jumped around a bit. Surprisingly, I got the zero to occur near the midpoint of the trimmer using an unmatched set of Chinese power tubes I used to start the amp with. However I could not get a zero (trimmer maxed out) using a matched set of JJs I installed. I swapped the tubes and then found some range on the trimmer to hit zero before the trimmer maxed out. The signal on the scope it kinda interesting. Here is the voltage on each 1ohm cathode resistor at 25 watts. I found the closest match when the scope was measuring amplitude on DC coupled input:

pi-balance.png

I was worried that the OD entrance divider might knock the voltage down too much. But I found that with it set on 50%, stage III was getting a 100mV signal. So, I think I will leave it alone for now. Below drive1 is on 4, OD level is on 3.5, channel 1 is the grid of stage III and channel 2 is the PI input.

od-in.png

Note that early breakup in the overdrive channel is not symmetric.

Finally I examined the FET trimmer. I found that at arund 50% the trimmer gave me a gain of about 5. My intended use for the FET is to juice up low output single coil pickups. A gain of five might be a little too hot. I'll wait and revisit this after playing the amp with my strat.

fet-io.png

[romberg]

Been playing it a bit and have noticed that it is a very versatile amp. In particular there are quite a few different sounding overdrive sounds you can produce. You can see these with a scope. Since I've not bumped into any pictures of an ODS on a scope I'll post a few shots here.

The first type of overdrive I can get is power tube clipping just driving the power amp with the normal channel. Anything over about 1.8Vrms into the PI starts to clip the power tubes. Since this happens with the normal channel drive on 5 and the level on 6, there is plenty of room to go to get the power tubes into heavy clipping.

Here I'll mention that I took these shots with the following amp settings:

Input: 100mVrms sine wave @ 1kHz (norm input)
Load: 16 ohm pure resistive
Rock/Jazz: Rock
brite: off
mid boost: off
HRM: on
TMB: 5
presence: 5
od trim: 5

Even more gain is available if the tone stacks are lifted. However it is probably more normal to have them in circuit.

If you increase the gain of the normal channel some more you can get the PI to start to clip:

pi-clip.png

c1=pi, c2=spk, nv=5, nl=8, mv=5

Here you can see the normal channel is still clean at around 4Vrms and the PI is starting to clip. PI clipping looks alot like (and sounds alot like) power tube clipping. Here you can see I've turned down the PPIMV a bit. I can get the sound of the clipping PI at very low overall volumes. However if one kept it cranked to 10 then you would have both power tube and PI clipping.

The normal channel all by itself can push the PI pretty hard.

pi-square.png

c1=pi, c2=spk, nv=5, nl-10, mv=5

Here the PI output is approaching squarewaves and the normal drive/volume is only on 5. Now dime it:

norm-max.png

c1=pi, c2=spk, nv=10, nl=10, mv=3

Here I think the output starts to become non-symmetric due to the second stage starting to distort. You can see that the PI signal is no longer a pure sine wave. This could be pushed further (as mentioned above) by engaging the PAB. But I did not go there. There is also a second way to give more punch with the normal channel. The FET input.

fet1.png

c1=2i, c2=pi, nv=6.2, nl=5

Here I'm putting the 100mVrms signal into the FET and moved the scope probes to either side of the second gain stage. With the drive on around 6 the second stage starts to distort. Let's crank it...

fet2.png

c1=2i, c2=pi, nv=10, nl=4

Here the second stage is looking square. But not the same shape (or sound) as the power tube or PI distortion. And sage 1 is starting to look a little distorted too. My FET trimmer is only set to 50%. So, no doubt one could probably get some distortion out of V1a if the FET gain was fully maxed out.

Next post I'll show some scope shots of the overdrive channel. The distortion here (not surprisingly) looks and sound different than what we see here.

Mike

Key

• nv = normal volume/drive
• ov = od volume/drive
• nl = normal level (master)
• ol = od level (master/ratio)
• mv = PPIMV
• c1 = channel 1 (yellow)
• c2 = channel 2 (blue)
• pi = phase inverter input
• spk = speaker out
• ni = input/grid of gain stage stage n

[romberg]

Finally the OD channel. Same settings as before. Now the OD relay is engaged.

v2-clip.png

c1=3i, c2=pi, nv=5, ov=4, ol=3.5

Here is the input to VIIa and the PI. The interesting thing is that the early bits of the overdrive channels distortion are non-symmetric clipping of VIIb. This of course has it own kind of sound. If we increase the OD drive some more we get:

v2-square.png

c1=3i, c2=pi, nv=5, ov=7, ol=3.5

At around 7 on the dial the channel's output is more symmetric. Note that the tops of the waves are getting a bit of a "dip" to them. This looks and sounds different from the distortion from the PI and power tubes. Increasing the OD drive to max:

d2-max.png

c1=3i, c2=pi, nv=5, ov=10, ol=3.5

The bottom of the square waves are now starting to get a bit of a dip in them too. The normal channel is only set on five. Let's make like Nigel Tuffnel and push it over the cliff:

all-max.png

c1=4i, c2=pi, nv=10, ov=10, ol=3.5

Now there is more dip in the bottom. Also the input to the fourth gain stage is no longer a clean sine wave. This is of course due to the second stage in the normal channel being distorted which we saw above. I also think that there may be signs that the third gain stage is also starting to distort. But I did not go down that road.

So, one can set up an amp like this to give layers of distortion. And the sound can change as a note decays (or feeds back) and the distortion from each sucessive layer fades or grows. No wonder these types of amps have a reputation for being tricky to dial in. There is quite alot going on inside!

Mike

[brewdude]

It’s very cool to see all the waveforms.
I would love to hear the various shapes...

Or, perhaps dial in a particularly satisfying Dumbly tone and see what it looks like and if you can glean any wisdom from shape and what was the necessary recipe (settings?).

Anyway, thanks.
Congratulations on an amazing build.