Paleo-Electronics

Clueless Idiots and High Voltage Vacuum Tubes Really Do Mix!

Paleo-Electronics header image 1

Songbird : Choosing The Output Tranny

June 5th, 2011 · calculations, Design, Songbird

Actually, it is already chosen. It is the existing 8k:8 center tapped primary that is already in the Crestline amp. However, I am excited by the choice, and at the risk of sounding like a dangerous lunatic, I’ll explain why.

First though, lets look at a standard SE output section.

Champ output section

Champ output section

That’s from a late 50′s (5F1) Champ, but they are all the same until the Medalist Skylark shows up. I’ve cleaned it up a little of the wires we don’t care about. So Vs is the highest voltage in the amp. It’s 305 to 315 in most of the old schematics and 375 for the 50′s Gibsons if the schematic is to be believed. Vs flows through the primary of the OT and into the plate of the output tube. Those are DC voltages (in theory) and in a perfect transformer, this would have zero loss and zero effect on the secondary.

So the DC voltage runs through the transformer to the plate. We are now past my understanding of the transformers action in one regard. Transformers with Vs through them are called energized and I’m not sure why this is done. With no signal, that is DC and there should be no energy in the magnet or coil. But amp guys always talk about an ‘energized’ transformer. I’ll dig into the RDH4 soon and maybe get an answer.

I would have thought that it would be better to capacitively couple the transformer, which I would have thought would result in lower requirements for the transformer.. but ok. What do I know?

As I said, in theory, this is a DC voltage but in fact it is only after the first filter capacitor. You might recall that there is significant ripple on the DC from the first filter (see here where I got 10 V pp on a 260VDC line). I wonder if that alone isn’t enough to create output hum.

So anyway.. signal comes in, gets amplified and shows up on the plate of the output tube. This varying current energizes the tranny which transfers energy to the speaker. The screen voltage is set to a value slightly lower than the plate for maximum output. And you must always have resistor (1k will do) on that screen.

However, the tube/tranny area is fertile ground for amp tweaking, primarily by lowering the screen voltage. There are a few ways to do this but I’ve bored you enough for now.

→ No CommentsTags:

SongBird : Output Tube(s)

June 3rd, 2011 · calculations, Design, Songbird

6AQ5

That was easy! Ok .. moving on.. Wait.. What’s that you say? You want to know why? Really?

Well, alright.. if you are that bored.
[Read more →]

→ No CommentsTags:

SongBird : Block Diagram

May 30th, 2011 · Design, Songbird

Getting started designing an Amp means making quite a few taste decisions up front. The only thing you really know about the amp is there is going to be a guitar plugged into it. The next few moves are up to you before you can start with circuit design.

The Bedroom Blaster.

My goal here is that creamy SE distortion at low bedroom volumes. I want low watt output, preferably even lower than the 5 Watt Skylarks and Champs. But maybe.. in the back of my mind.. I might want to boost that a bit sometimes. I also want a lot of flexibility in the amp topology. Basically, anything that can be switched, will.


Here is what most small SE amps look like :

Basic SE Topology

Basic SE Topology

Input to preamp to power amp to speaker. Just about as basic as it gets. Very little changes between the various SE amps (at least the ones that I fool with).

Amp Preamp
Tube
Output
Tube
Notes
Champ 5C1 6SJ7 6V6 Pentode Preamp
Champ 5E1 12AX7 6V6
Champ 5F1 12AX7 6V6
Gibson Skylark (50′s) 12AX7 6V6
Gibson Skylark (Medalist) 12AX7 6BQ5



Because I want to drive it hard, and I know a bit about SE amp topologies, I know that I want to add a second preamp stage to my amp. A single 12AX7 with its two triodes is easily capable of blasting past the limits of a 6V6 output tube, so we know we will have to find a way to contain two full stages.

Now if I want to double the output, I could, in theory have a second output tube in parallel and maintain Class A1 amplification. This is not the same as Push-Pull, which most two tube amps do, and is Class AB1 (usually). I think the GA-6 ‘Gibsonette’ used parallel output tubes in Class A SE config.

Here is a slightly expanded block diagram :

Songbird Block Diagram

Songbird Block Diagram

No decisions yet about that second output tube, but I have an extra tube socket..

→ No CommentsTags:

SongBird : Time To Get Busy

May 29th, 2011 · Design, Songbird

So that Medalist is working ok. Sound is a bit thin for lead work but that’s mostly because it is still too loud for me to crank it at full. Beautiful sounding amp if you pickups can put out some sound or if you boost leading in to the amp.

So now it’s time to design my dream SE bedroom amp. I’ve mentioned decisions and design considerations before, and following the songbird category will let you slog through all that if you like.

I’ve changed my mind on a few things and come up with some ideas. I’ll sketch here and some of these will be explored at length later.
[Read more →]

→ No CommentsTags:

Sidebar : Graphing Resistances

May 22nd, 2011 · Amp Repair, calculations, Datasheets, Schematics

I want to talk about the power section of my Medalist in detail (and some other things), but before I do that I’m going to give myself a refresher course on tube data sheets and graphing resistances.

Recall, if you will, that V = I * R.  An equation in three variables.  If any one of them is held fixed, then you have a simple linear equation.  For example if we held R constant at say 4600 Ω, then we would have V = 4600 * I.    For giggles, I will choose to make I a function of V,  so I(V) = V / 4600. I choose a value for V and the equation tells me I. Lets look at a picture. Here we see resistances graphed for 500, 1000 and 5000Ω. Just like any other graph, you pick a value on the x axis, go straight up for to which ever resistance line you like, and then look across to read off the current through that resistance for the given voltage.

Simple graph of Various Resistances

Simple graph of Various Resistances

That’s all well and good. The key point is that resistance is the slope of the line. The only thing is, in general, we graph voltage drops across resistors.

Circuit that needs graphing

Circuit that needs graphing

In the example above, the bottom part of the circuit is likely connected to the plate of a tube or something and we frequently need to plot how the current changes the voltage across the resistor (thus telling us the actual plate voltage).

In order to graph a linear equation, we need two points. Let us calculate two simple points using v = i*R. We know R. We pick i = 0. No current flowing. With no current flowing, there is no drop in voltage so the voltage at the bottom of our little circuit is the same as that as the supply (200V). We can graph V=200, i=0 and that is one point. For the second point, we choose V=0, which means that all of the voltage has dropped through the resistor. i = V/R = 200/100k = 2mA. We can now draw a line between those two points and read any values we need from there.

Graphing Loads

Graphing Loads

Notice I also converted to milliamps.   By now it is surely obvious why I did all this, I want to start reading data sheets, at night.. to my kids.

→ No CommentsTags:

Medalist Debug – A Test Drive

May 4th, 2011 · Amp Repair, Medalist Debug, sounds

So, I’m still working on posts for the calculations and all that and I’m still uncomfortable with the running voltages but it seems to be doing what it was designed to do, so hell, let’s hear it!

67Skylark

67Skylark

As luck would have it, I got a genuine guitar legend to try the thing out! Johnny “V” Vernazza (http://johnnyv.net) can be heard hammering it on my
YouTube clip here. Sadly, this cuts off early, but the full five minute MP3 is here : JohnnyV puts the medalist through its paces

So $175 with shipping off ebay and $30 or so in parts and I have a pure Class A, SE amp that simply rocks.

I’ll post more on the engineering side soon, but I thought I should make it clear that whether or not I understand it, it sounds fantastic.

Also.. you really should think about lessons from Johnny V if you live in the So Cal OC/SD area.

→ No CommentsTags:

Debugging the Medalist – More Weirdness

April 23rd, 2011 · Amp Repair, calculations, Medalist Debug

So I replaced the bad resistors and took so measurements. Basically, everything looks perfectly reasonable, and it sounds great at low volumes, but I have one major problem. Anywhere past 3 or 4 on the volume knob slams the output tube with up to 30 V (rms) on the grid! For reasons I may show, the output tube grid should not be more than 7 volts or so.

The real problem now is that every thing seems to be running to spec (other than that wacky grid problem). Lets break it down.
[Read more →]

→ No CommentsTags:

Debugging The Medalist – The Output Tranny

April 7th, 2011 · Amp Repair, calculations, Medalist Debug

Well, the parts have arrived, but first, I want to understand that output section a bit better so let us examine the OT.

The always helpful Patrick at Mercury Magnetics suggested that the drop in replacement they make is the GA5-0 which has an 8kΩ primary to 8Ω secondary. He also noted that they varied widely, with little by way of information about why or when. We here at Paleoelectronics [snicker] are familiar with that.
[Read more →]

→ 1 CommentTags:

Debugging the Medalist – Mistakes Were Made

April 7th, 2011 · Uncategorized

As I try to figure out what is wrong without blowing more 6BQ5′s to do it, I have to acknowledge that I rushed a couple of things in my excitement. I guess I’ll never really be a radio man.

Whether or not those mistakes caused the failure is not likely to be known. However, those mistakes did cost me in valuable information. If I had gone through the amp a bit more thoroughly, it is possible I would have detected a condition that caused the blowout. It’s very possible that I would have avoided blowing the resistors since I would likely have not run the amp (especially dimed) and it is certain I would have more information now to work with.

What mistakes?

  • I did not measure grid voltages. Especially in the output tube, where the plate and cathode values are fine, the grid input could well have been the problem if indeed the problem was external to the tube.
  • I did not figure out that I knew what the design voltages were until after ( I should have checked the MSM schematic against the one I got. )
  • I did not scope a clean signal through the amp.

These wants shall be rectified.

“Live and don’t learn, I always say!” – Calvin

→ No CommentsTags:

Debugging the Medalist – Intelligent Design

April 5th, 2011 · Amp Repair, calculations, Medalist Debug

As we now have some reference voltages to use (see previous post) we can take a lot of the guesswork out of our debugging. So here is the power section from the Master Service Manual.

Those are design values. The test readings are in the top corner of that page (319 in the MSM) and very hard to read. Here is the critical point, that 260V goes through the output transformer and loses a bit before it gets to the plate. Lets presume it drops 10V because it looks like the test reading on the plate is 250V. The 250V source off the 3.3KΩ goes directly to the screen. Those are exactly the GE design center values for this tube. 250/250 and -7.3 grid. These numbers make perfect sense. Most engineers would have used the design center values unless they were forced to change.

Lets see how things looked before the blow out.

Test Point Name Value Units Expected Notes
TP1 PT 2nd 255 VAC 260 Expected is somewhat dubious, very hard to
TP2 PT 2nd 255 VAC 260 read on the schematic
TP3 Rect. Out 284 VDC 260 A bit hot here, probably my AC is hot.
TP4 Heater 6.43 VAC Heater
TP5 POW Screen 262 VDC 250
TP5.5 POW Plate 263.6 VDC Not Given. SB slightly less than TP3
TP6 POW Cathode 7.79 VDC Not readable.



Pretty decent agreement really. I’m running a bit hot, but well within the usual 10% slop that should be tolerated. So I still don’t know what happened, but I am really starting to believe the mechanical failure theory. Parts arrive in a couple of days.. maybe I will get in a power section analysis before then.

→ No CommentsTags: