4-65a SE Amp: 46 filament supply tested

Finished the second channel and tested the filament supply. The filament array which is formed by two paralleled pairs of 20W 10Ω wirewound resistors gets hot as expected. The array temperature is about 110-126°C at an ambient temperature of 24°C. The anode of the 46 gets to 49°C after 20 min of use and the heatsink stays at 42°C whereas the regulator TO-220 transistors are about 45°C. There is about 30W dissipated on each array. Yes I know, a lot of power but the filament bias is hard to beat in terms of sound in my view.

4-65a SE Amp building process

Some drilling and mounting work done this morning on the 4-65a SE amp breadboard:

Filament bias resistor array

Closer view of the filament bias resistor array

Mounting pilars for the SSHV2 regulators

46 valve sockets

Closer view of the 46 valve sockets

46 sockets, filament bias array and filament and HV supply connectors

During the afternoon, I managed to wire a filament supply for one channel 46 driver. Tested and working ok, now can move to the next one:

Filament regulator with its heat-sink at the back, anode loads and the 46 DHT drivers at the centre.

46 DHT driver under test. Gyrator anode load can be seen at the front

46 DHT driver

4-65a SE Amplifier: testing filament regulators

The Rod Coleman Filament Regulators were already tested in the New Year. It seems a long time indeed. Here is a selection of pictures after fitting the regulators in their heat-sinks and doing a final test before assembly:

4-65a

Testing the 4-65a filament regulators

4-65a filaments are lovely

46 filament bias and the regulator

46 Driver regulator

4-65a SE Amp: Shunt regulator

Thought it was going to be an easy task as I’ve done it before many times and building a Shunt regulator seems to be not the challenging part of this amplifier build. We all know that life brings surprises and specially when we are not expecting them. My 4-65a SE amplifier requires a very stable DC as part of the DC-coupling design. The Salas Shunt Regulator version 2 (a.k.a. SSHV2) is a good choice for this task.

New pass FET

Fitting new pass MOSFET and CCS

Salas SSHV2 shunt regulator

After building it very quickly I struggled to get it to work. To cut a long story short which involved some IRF840, PNP and JFET replacements, I discovered that the stabilising RC wasn’t connected as the 330nF MKP capacitor was not properly soldered to the right holes. The PCB has multiple holes to accomodate capacitor sizes, however only the top two correspond to one capacitor pin and the remaining bottom ones are for the other. My logic of placing the capacitor in the centre clearly didn’t work and the capacitor was disconnected in the end. Finally, when hooking the regulator to the raw supply and switching it on, the whole thing produced the unwanted smoke particular of sand devices getting blasted. What happened? The maximum input voltage to the regulator evidently exceeded the CCS voltage and the top FET (M1) blowed away and therefore the regulator cascode CCS (J1) and the pass FET (M3) as well. My PCB was already suffering from multiple solder work and was reaching to its usable life. I looked at using HV parts as hand to increase the robustness of the regulator. The pass-FET was replaced by a 1kV part (STE5NK100Z) and the Mosfet CCS DN2540 pair for an IXTP01N100D which is also 1kV part:

All worked well until I realised that the differences between DN2540 and 01N100D’s VGS(th) and gm made the CCS maximum to be limited to about 40mA given the test point resistor value. As M2 can be a simple DN2540, I replaced it back and all worked well to get 60mA and deliver about 280V @ 40mA rock-solid!

4-65a SE Amp: HV3 supply

Building the HV3 (+330V Supply)

I could easily say that by now I’m tired of building power supplies. Yes, I’m and fundamentally can’t see the day when I get to fire up this amplifier.

Filament supplies and three stacked power supplies is the price I’m paying to get a completely cap-less and DC coupled A2 amplifier. I guess that my analysis once completed will be made with full perspective of every single implication of this amp: iron, heat and weight. Yes sir, this is a heavy-weight challenger.

I guess that this specific HT power supply design is quite flexible as could be easily reused in many of the projects I have in mind, which unfortunately keep growing.

Good thing is though, I can make these supplies pretty quickly, but don’t do this at home ok? There is serious HT involved. I don’t have pets or kids (but do have a wife) and these supplies should be hidden and away from any poking curious finger.

The final design is similar as the one used before. Could be adapted to choke input, but with the components I had at hand, this supply is very well filtered. It provides only 15mV ripple noise which at 330V is a lovely noise floor around -87dB:

The mains transformers are Weiss (excellent quality) which has screened windings. Instead of having a full Graetz valve rectifier and waste more heat (and use more damper valves), the rectifier is hybrid using a pair of UF4007. Capacitors are oil and polyester ones for the output HF decoupling on each rail. Each channel will feed a Salas SSHV2 shunt regulator which will provide the stable DC reference for the 46 driver stage and bias point for the output 4-65a stage.

Bad news is that, there is one supply left to be built before I can test at least one channel!

4-65a SE Amp: testing the 600V raw supply

Burning the 6D22S damper valves

After a failed initial test which caused the silicon and the damping series resistor of the hybrid bridge to blow up due to a gassy NOS 6D22S, I burned in the valves for 30min and tested the supply with a dummy load.

Some few changes to the supply design included the replacement of the HV diodes with a series pair of 1200V@5A fast diodes. The filaments were referenced to cathode (+B) instead of ground to minimise Vhk and a snubber network formed of a pair of 100nF LCR 1.5kV capacitors across the input choke were added.

Supply under test

This is a hefty supply: massively heavy. The heaviest piece of iron I have built so far. Anyway, great decision to build this amp in a modular fashion, otherwise I would have need some mechanical aid for moving the stuff around and clearly my wife would have kicked me out of the house (she hasn’t seen this yet so the latter is yet to happen).

The load resistor is a pair of 3k3 50W aluminium clad ones bolted to an aluminium piece which at the same time is held by a vice (thanks Rod for the suggestion). This provides sufficient mass and heat dissipation capacity (need about 120W).

Initial tests were great, need to do some further ones.

here is the diagram so far:

Update 9th Feb 2013: Minor updates including rightsizing of the mains fuse and removal of earth switch as this supply will be used in floating mode

4P1L pentode driver (continued)

After the initial tests done with the 4P1L in pentode mode and filament bias, I thought a bit about how this driver could be implemented in practice. I may try this configuration in my 4-65a SE amp, but am not urged by this at all.

The screen supply is formed by a gas valve (SG3S) which provides a very stable reference when feed by a CCS. In this case the cascoded pair M3 and M4 will provide in conjunction with R4 and U2 a very low noise screen current to U1. R9 has to be adjusted on test to set a current of about 15mA on the CCS. The 4P1L will draw 1.8mA at 81V as screen current, so R4 may also need adjustment to set the right operating point.

The driver should provide about a gain of 150. Driving easily a transmiting vale (or why not a 300B 🙂 ) in class A2 with this configuration. My tests showed a maximum THD of about 0.27% at 200Vpp.

Gain could be reduced if needed by tweaking the RL. And if stacked supplies are not used, then a single +400V supply could be used with additional dissipation across the reference currents (M1 and M3/M4).

Keep thinking…

46 DHT driver final tests

Having built the 4P1L filament bias driver stage in a breadboard, I now have the sufficient voltage swing to drive the 46 to maximum sweep. In my 4-65a SE amp, a maximum of 200Vpp is required to drive the amp into class A2.

The following tests conditions were used:

4P1L first stage:
- DN2540 gyrator in mu follower output
- 220nF/450V Capacitor coupled into 46 driver
- Filament bias: 15 ohms, Vgk=-10V
- Vsupply=355V and Va0=210V
- Output set to about 30-32Vpp to drive 46 at 200Vpp
46 driver stage:
- IXYS 01N100 gyrator in mu follower output
- Load impedance is 100K (Pete Millett’s interface)
- Filament bias: 10 ohm / 100W Vgk=-17V
- Vsupply=355V and Va0=204-208V
- Output set to 200Vpp

I tested 28 valves. Just a few of my lot are NOS. The average THD was about 0.4-0.5% but a good selection of 8 valves (mainly Sylvania NOS) provided a consistent 0.18% THD:

Happy now with the initial tests and selection of 46 pairs for the amplifier, I can now continue with the build…

4-65a SE Amp: refining the 46 driver

I did some tests today and looked at minimising distortion of this 46 driver in filament bias and found that Va=230V (instead of 184V) to provide best performance:Filament bias resistor array is now laid out horizontally to improve the dissipation of heat.

Here is the performance (0.05% at 17Vrms) at maximum drive input from my audio test set:

4-65a SE Amp: testing the 46 driver

46 driver breadboarded. The mu-follower gyrator, the filament bias resistor array and the nice teflon UX5 socket from Jakeband. The filament bias resistor array is formed by 4 10Ω 20W dale wirewound resistors. These get very hot so probably need to think an alternative layout or further resistors in parallel:The performance is very good. I just picked up a random 46 from my stock and biased it at 204V (which is the operating point in my design) achieving less than 0.05% at 10Vrms. Need to re-run this test to see how will perform at 70Vrms:

Using my spice model created from a good 46 valve, THD should be around 0.15% at 200Vpp with a 100K load and performance is great to loads down to 100 ohms. Clearly the load in A2 will change from high impedance to some kΩ so this driver should maintain outstanding linearity all the way through: