Current limit resistors for GZ34?

[martin manning]

...Yes, I got it 40 ohms per leg, I wonder if that will effect tone of the amp?

The desired rail voltage is 320vdc, so with a secondary of 303 - 0 -303 I should have a bit of wiggle room.

The power dissipated should be a follows.

P = (1.1 × Idc)^2 × R
= (1.1 x 0.250)^2 x 40
= 3.025 watts

Another 40Ω at 250mA would drop 10V, but you will get more that that, I think. Vpeak from 303 VAC at 428, so you should have room. I'd make the 40Ω resistors at least 5W, and higher would be better to keep the surface temperature down.

[Mark]

Thanks Martin.

[R.G.]

Just to insert a slightly different approach...

I posted this http://www.geofex.com/Article_Folders/m ... 0Clamp.pdf
at geofex back in 2012. It uses a power MOSFET to watch the current in the wire between the rectifiers and the first filter cap. The MOSFET stays low resistance until the current in R2 exceeds the conduction voltage for the base emitter of bipolar transistor Q2. When that happens, Q2 steals the charge from the MOSFET's gate and the MOSFET starts being higher and higher resistance. In operation, this effectively means the current can rise to some level, then is hard-limited by the MOSFET not letting more current through.

The MOSFET, like any resistors put in to limit current in the rectifiers, needs to dissipate some heat, about 5-10W like the resistors would, and is floating up on top of the B+ voltage. The article shows it on the positive side of the rectification. If you have properly wired your CT only to the negative side of the first filter cap, this same circuit could be inserted into the wire from the CT to the first filter cap and work properly sitting down around ground - as could a limiting resistor for the rectifier tube.

[martin manning]

...If you have properly wired your CT only to the negative side of the first filter cap, this same circuit could be inserted into the wire from the CT to the first filter cap and work properly sitting down around ground - as could a limiting resistor for the rectifier tube.

Interesting idea, and simple to implement in a safe way: ground the CT through this gadget, and ground the reservoir at the same location. No special treatment is needed since the drain is grounded.

[SoulFetish]

...If you have properly wired your CT only to the negative side of the first filter cap, this same circuit could be inserted into the wire from the CT to the first filter cap and work properly sitting down around ground - as could a limiting resistor for the rectifier tube.

Interesting idea, and simple to implement in a safe way: ground the CT through this gadget, and ground the reservoir at the same location. No special treatment is needed since the drain is grounded.

So, Martin, you are talking about configuring this as follows?

Mosfet_Current_Clamp_w-Drain_Grounded.jpg

Well, since the current going in, is equal to the current returning, is there any reason this wouldn't work as well?

[martin manning]

Yes correct, and as per RG’s suggestion.

[Colossal]

Mark said:

I have a Mercury Magnetics Haddon clone transformer and the figures are as follows:

Rpri = 3.3 ohms
Rsec = 28.2 ohms on one side and 29.5 ohmson the other side.
Vpri = 240vac
Vsec = 303.5vac

Here is a typical spec sheet for a GZ-34. The spec sheet gives a different equation for the series resistance.

Rt = Rs + n^2*Rp

...and Pete responded:

...As stated before the power transformer has a 303 - 0 - 303vac secondary...

That's crazy high, useless for an AC30 IMO.
I don't see the point in starting this unless you can get a more suitable PT.
The Woden in mine puts out 586Vac with the HT unloaded, 560Vac at idle, 560Vac at max 'clean' 25W, 554Vac at fully overdriven squarewave 38W. The HT idles at 317Vdc.
243Vac mains feeding 245V PT primary, heaters 6.41Vac at the EL84s.
Measurements taken with calibration checked Fluke 189.

I guess that if you've already got that PT, yes you could try bringing the voltage down, either with ~100ohm 10W in series with each HT winding phase, or a suitable zener, but that's adding yet more heat to an already hot amp.

I've been doing a bit of number crunching this weekend, fitting a model to look at the way Mark's PT is supposed to behave. On the surface, Pete's comments are right on...That PT at 303-0-303 seems like the rectified DC would be excessive. I am also looking at a different Haddon PT which is supposed to give a very low DC under load. I would expect both PTs to be "in the ballpark" for AC30s.

Ok, so the datasheet for a GZ34 shows an expected rectified voltage of 330VDC with Mark's setup (rounding 303 down to 300VAC...) and ensuring the secondary resistance adds up to 75Ω:

GZ34.png

However, when I use the parameters Mark listed for his Haddon PT, and 75Ω as the secondary resistance for the PT plus added resistance for the tube rectifier to make up the balance, I get a much higher DC for B+1 under load. Under the higher B+ the total EL84 current is reduced to 143mA (I am basing this on MaxDiss of 12W/B+).

EDIT: REMOVED FOR CORRECTION

Now, when I look at a PT I've been considering using, the secondary is 285VAC and I would expect it to perform very similarly to your Woden, Pete. I have it on good authority that under load, it should give a mere 295-300VDC or so with a GZ34. This implies a rectification factor of just 1.04; quite saggy. But to get the model to yield this low rectified DC voltage under load, the secondary resistance has to be increased to 400Ω. This seems really high and has me scratching my head! But this might make good sense because it limits the current inrush and ripple current and actually gives quite decent regulation. At these lower voltages the total power tube current is increased to 182mA (again 12W/295V B+). Also note the big difference in ripple current because of this high secondary resistance. With 32uF as the reservoir cap, the ripple voltage is 12%.

EDIT: REMOVED FOR CORRECTION

So I am a bit at odds with this 400Ω. I am reasonably confident my calculations are correct (scratches head...). I've probably given myself enough rope here to hang

[Tony Bones]

Tried it on PSUD and found that it requires 150 ohms to give 292VDC @ 195mA.

edit: changed the load slightly to get 296VDC @ 184mA. Closer to your numbers. Resistance is still 150 ohms.

[Colossal]

Tried it on PSUD and found that it requires 150 ohms to give 292VDC @ 195mA.

edit: changed the load slightly to get 296VDC @ 184mA. Closer to your numbers. Resistance is still 150 ohms.

Thanks a lot, Tony. I appreciate you running that on PSUD. I had taken the load resistance as VDCpk/total load current, however, I must not be accounting for the rectifier voltage drop correctly. Your simulation is showing a load resistance of 1.6kΩ and setting my secondary resistance to 150Ω, I would still have to subtract some excessively large voltage (for the rectifier) from rectified VDCpk to bring my 2.0kΩ load down to 1.6kΩ. A GZ34 should drop something like 17-20VDC, right? Hmmm, something is amiss...

EDIT: Ok, If I take ((VDCpk-GZ34 Drop)/(total current)) x 0.707, then BOOM, the whole thing drops into place

PSrev1.png

[Tony Bones]

With that much resistance (even 150 is a lot) the B+ becomes very sensitive to load current.

The GZ34 is not a linear resistance (not "ohmic") but it will drop around 15V at 185-200mA, so you might simulate it with a 75 ohm resistor. One for the whole B+ or one in each diode leg. Either way.

[Colossal]

With that much resistance (even 150 is a lot) the B+ becomes very sensitive to load current.

Yes, I've noticed that!

Ok, so at least it appears my model is behaving itself and I got the source of error straightened out. I appreciate you running the numbers in parallel. I will have the transformer in hand fairly soon and will set it up to test these assumptions.

The GZ34 is not a linear resistance (not "ohmic") but it will drop around 15V at 185-200mA, so you might simulate it with a 75 ohm resistor. One for the whole B+ or one in each diode leg. Either way.

I will do that. I was using 17V as the drop just wanted to make sure my assumptions are reasonable before going around pontificating in Mark's thread.