Dynaco Mark III Classroom

[peterh]

The risk with removing output stage tubes, and then powering up, is that the filter caps and coupling caps may well be exposed to noticeably higher voltages than those caps are rated for. That can be exacerbated by having a modern-day higher mains AC voltage than the vintage amp was nominally set up for. The main filter caps may also need a soft start to allow reforming and lower the initial leakage current to a survivable level. It is not unheard of for a bad main filter cap to take out the power transformer and/or the rectifier tube. Likewise the rectifier tube could be gassy and start arcing and it is not unheard of for that to then take out the power transformer.

That is in the picture. My recommendation in my first post should maybe mention that the bias test should be
brief ( less then 15 seconds). That's also why i mentioned to measure the remaining voltage on the B+, if
it's still around 200V then the cap is not fatally broken.

 

[gadget73]

You can also remove the rectifier tube and make this a non-concern. No high voltage, but the bias supply will work just fine.

 

[peterh]

You can also remove the rectifier tube and make this a non-concern. No high voltage, but the bias supply will work just fine.

Yes, that's an even better idea. Then when bias is checked reinstall the rectifier and make a short run observing the B+.

 

[Donstaff]

Whow! Life in an aging tube amplifier can be vicious. After reading this far, my inclination is to take Trobbin's advice as stated above:

"Different paths are taken by different people. Some will go straight for the bulb-limiter and turn it on. I go completely to the other end of the spectrum and quickly test all parts in a step by step manner, and appraise mains AC safety issues, and energise sections in a step by step manner, because I can do such tests as I have suitable equipment (eg. a variac and a megger) and experience. That ends up with basically a safe amp that has DC voltage levels around all of the circuitry, and before any signal is applied. That often also requires more than one set of valves, or other valve testing gear, to have confidence that valves don't have leaky grids, and can sit the anode voltage at a reasonable level so as to bias the PI stage appropriately, or balance the output stage cathode currents. So I would suggest that there can be a lot to do and learn before the amp even sees an oscillator on the input."

But there are a few things that I still do not understand about Trobbin's advice.

1. Does reference to the "bulb limiter" mean placing an incandescent light bulb in series with the voltage supply to the amp when first powered up? I need to learn how to do that.

2. Not having the suitable test equipment to do Trobbin's recommended step by step approach, would using the bulb limiter method with a 50 ohm load resistor across the speaker terminals, with the 6550 tubes removed be safe for the questionably sound or possibly leaky valve grids or capacitors?

3. What does "...sit the anode voltage at a reasonable level so as to bias the PI stage appropriately, or balance the output stage cathode currents." mean?

Don

 

[Donstaff]

So I have to have a variac to get started without endangering any components?

Don

 

[peterh]

So I have to have a variac to get started without endangering any components?

Don

No. It is good to have but not necessary.
Check that the fuse is according to manual ( 3A )
If you follow my plan , first check bias voltage on pin 5 with powertubes and rectifier removed. Turn bias pots to maximum ( most negative) voltage.
Then install rectifier and for a short while turn on, watching the B+. Do this several time with some minutes
between until you have 475 volt on b+.
Then install powertubes and a load resistor or speaker. Adjust bias and check that the can cap stays cool.

 

[Donstaff]

Thanks patterh. That is extremely helpful,

Don

 

[trobbins]

Many of us have a light bulb limiter jig, but they do require mains AC wiring and introduce a benchtop hazard, and require some learning curve as to what they do and how to use them effectively - lots of articles and threads on the web. By itself, it may not help reform e-caps that have been dormant for 30+ years - that may need an additional temporary work-around to be assessed (such as inserting a resistor in series with the B+ feed or using a suitably low-wattage bulb). Starting with low wattage bulb and working up can mitigate the consequences of any fault, and as described by peterh and gadget73, removing valves and then step by step inserting them can be used to confirm separate parts are functioning in a low risk manner, with the proviso about capacitor rated voltages being exceeded (which can be checked at the start by measurement of the heater voltage). If you had a nominal 240Vac mains, then another option instead of a variac is to use a 240 to 120V step-down transformer, as they can be common and often come in a suitable commercial form (appropriate plugs and sockets and protective earth).

Peterh's technique of energising B+ in short duration squirts is I believe aimed at using the valve diode's slow thermal start up to limit B+ feed current in to any fault or cap needing reforming or as a means to measure circuit voltages in time to turn the amp off if voltages are exceeded (eg. you have a high mains ACV issue). That could be acceptable - although capacitor manufacturers recommend including a form of current limiting (eg. a series 1k resistor is often identified in recommendations) for at least 30 minutes.

Some people often observe a weak rectifier valve arcing during warm-up periods - the typical forum recommendation is to bin the rectifier if that happens, as the arcing indicates the cathode has been damaged. Including 2x 1N4007 diodes in series with each GZ34 anode (given a 430-0-430V secondary) alleviates the chance of the GZ34 arcing, but without testing the PIV performance of the GZ34 then you don't know if you have a poor GZ34 for starters, which may need changing as the valve may cause increased ripple voltage/hum).

 

[Donstaff]

Thanks robbins. I'm learning a lot and it is great to have you and others so freely sharing your wealth of experience. Got my new bottom cover today and hope to get to take a look at the Mark III next week.

Don

 

[Mr Boochie]

I suggest the other way around. Use a 1W 10R resistor for tying up the output. Apply no signal; that is for when you have some real dissipation available in the load. The 1W element is to give you notice something is wrong before it tears something expensive up. If the amp oscillates, you may smell the hot R in time to shut off the amp and troubleshoot at length. If you wish to get really creative, wrap the resistor in a wee strip ov bacon...you will know exactly when the amp is in trouble...

As to tracing the signal...I like to use wee liddle arrows for increasing signal or decreasing signal( the part of the waveform ). We are looking at an instant...and go from there, assuming the whole bit of the amp is at that same instant.

Example: the grid 1 of the 6AN8A, draw a little arrow pointing UP. move to the plate, little arrow going down....off towards the triode section. its grid going down, cathode going down, and its plate going up. So we now have a + and - signal at the plate and cathode of the 6AN8's triode. This is the bit called 'phase splitter', or phase inverter. Its job is to convert SE to PP signals...these get delivered to the output tube's grid 1...and you can repeat the arrow-ing process.

Cheers,
Douglas

I use a 5-watt wire wound resistor load. For the same reason. At 100-watts, it'll be a battle between what goes first. The $6 resistor or the $200 output transformer.

 

[Schmidlapper]

I use a 5-watt wire wound resistor load. For the same reason. At 100-watts, it'll be a battle between what goes first. The $6 resistor or the $200 output transformer.

View attachment 2125580

I suspect you also have larger dummy loads for performance testing. If a severely under rated dummy load explodes under load during test, it really wouldn’t mean the amp has an issue. I understand suggesting using a low rated dummy load as a fuse initially, but at some point it is prudent to switch to a larger one. There are even tests where no dummy load is appropriate.

 

[Pio1980]

I suspect you also have larger dummy loads for performance testing. If a severely under rated dummy load explodes under load during test, it really wouldn’t mean the amp has an issue. I understand suggesting using a low rated dummy load as a fuse initially, but at some point it is prudent to switch to a larger one. There are even tests where no dummy load is appropriate.

https://www.parts-express.com/8-Ohm...l24VvmBL_DhKvveyPaIty-z05zkuVEx8aAi-sEALw_wcB

 

[Mr Boochie]

I suspect you also have larger dummy loads for performance testing. If a severely under rated dummy load explodes under load during test, it really wouldn’t mean the amp has an issue. I understand suggesting using a low rated dummy load as a fuse initially, but at some point it is prudent to switch to a larger one. There are even tests where no dummy load is appropriate.

No, for high output tests I use an old 12-inch speaker. Facing the floor.

 

[Schmidlapper]

That could easily be a safer rated dummy load. If an under rated dummy load fails the output transformer becomes the new one instantly.

 

[gadget73]

Speakers don't make a great high power dummy load. They have non-simple impedance properties. If you want to do a frequency response test, it relies on a constant known load so your measurements make sense.

Also my ears have ringing problems, I'm not doing full power testing on anything and then having to suffer the high pitch screaming all night that will result from such silliness.

 

[Donstaff]

I am looking at the Mark III now and recalling the work that I did on it nearly 20 years ago. Turns out that to the best of my recollection, I somehow at the time determined that the can capacitor needed replacing. I replaced it with a CDE DD0179.5A, which had four sections of 15 to 20 MFD 500 W.V. it now looks like that capacitor has a voltage rating that is too low.

The dynakit company's multi-section can capacitors which they say are suitable replacements for the Dynaco MK III amplifier are described as the
23222-6, with the specifications:

30/20/20/20 uF @ 550/600 VDC FP (twist-lock) can style. (1) 30 uF @ 550 VDC (600V) Surge (3) 20 uF @ 550 VDC (600V) Surge Tolerance -10/+30%, Temp Rating 70 deg C 1 3/8″ dia. x 3 1/2″ ht. (from…



The company also offers another multi-section can capacitor #298906 with the specifications

30/20/20/20 uf@ 525 vdc fp (twist-lock) can style. (1) 30 uf@525 vdc (3) 20 uf@525 vdc 1 3/8″ dia. x 3 9/16″ht (from base) Fits: DYNAKIT ST-70, MK III, MK IV Note: Caution, voltages with MK III may exceed 525…

The (apparently capacitor) capacitor in my Mark III has specifications of: 20 to 30 MFD 525 W.V., 15 to 20 MFD 525 W.V., 15 to 20 MFD 525 W.V., and15 to 20 MFD 525 W.V.

I wonder if there is a history of abnormal failure of the original smaller can capacitors, and if the Daynahit 30/20/20/20 uf@ 525 vdc capacitor is adequate or should I go for the $26 more expensive version? Also, how would I test the old original can capacitor which is now out of the circuit?

Don

 

[Donstaff]

I suspect you also have larger dummy loads for performance testing. If a severely under rated dummy load explodes under load during test, it really wouldn’t mean the amp has an issue. I understand suggesting using a low rated dummy load as a fuse initially, but at some point it is prudent to switch to a larger one. There are even tests where no dummy load is appropriate.

 

[Donstaff]

I ordered a cheap 100 W resistors, but after seeing the one with the leads, I ordered it too. Looks like good insurance against an output transformer becoming a $200 fuse!

Don

 

[trobbins]

.... how would I test the old original can capacitor which is now out of the circuit?

I'd recommend a vintage e-cap is tested for capacitance (using something like a cheapy multimeter capacitance meter), and for leakage current.

The leakage current test can be done in a number of ways, depending on the equipment you have. The voltage supply is preferably variable, and up to the rated voltage of the cap, and a series resistor like 1kohm is inserted in series, and a DMM measuring the voltage across the 1k is used to tell the charging/leakage current. A capacitor that hasn't been used for years needs some care to allow it to form up its insulation without passing high levels of current - that is typically done by increasing the voltage supply slowly or in steps, and waiting for the charging current to subside. Then when the supply is up to rated voltage, leave it on for half to 1 hour and note the leakage current after an hour.. If you have multiple caps, then compare the leakage currents. Do this test with caution in case the capacitor fails (the voltage across the 1k will increase). An insulation resistance tester set at 500Vdc would also achieve the same outcome (although such testers are not designed for anything but short operation). The original amp can be used as the voltage supply, by using a variac input, and a solid-state diode replacing the valve diode, and disconnecting all other circuitry.

Some people also have ESR meters, which can provide a guide, but are not an absolute measurement unless you have tested a significant population of the same capacitor, at the same temperature, and use population statistics to know if the next sample cap being tested is 'in the bunch' or an outlier.

 

[Donstaff]

Thanks Trobbins. Any comment on the first part of my question about the first two can capacitors', Dynakit's 23222-6 vs. their 298906's suitability for the Dynaco Mark III?