Search results

124 bit… shocking! (I used to sell optical quantum cryptography before we got down that rabbit hole).

Killed the oscillation by rewiring and removing a diode. Currently mosfet sees a 37Vds at 62mA which means 2.4W of heat, which is above what I'd spec'd for. As a result the thermal increase causes a slow drop of current/voltage over time. What it should really be is about 4Vds and about 0.24W...

Could you not cascode them?

Hmm. Just modelled the regulator this morning in ltspice and it seems to have an oscillation at 1MHz. It looks like the threshold of the mosfet is triggering it, then off and then on. I'll add the 100K into the RSET path which should work in practice but given the model is throwing its toys, it...

Ok, so the mods have been made to the maida regulator. WIth 122Vdc from the bench supplies (120mVpp noise) and a 5k5 resistor providing a 28mA load, It's about 63mV noise at the moment but given the noise increases when I attempt to save a scope image, I think we're on the boundary of the...

So I'm now down in the weeds (for tube amps) looking at the 10mV to 500mV range, and there's a good reason - my linear power supply is simple as it's a toroidal -> bridge rectifier -> RCRC (390uF each) -> B+. The B+ is reading 340.33Vdc but it has between 0.2-0.3V of ripple still according the...

I've just been doing some ESR calculations for my tube compressor as the Vishay 0.1uF has an ESR at 62Hz of 7R7! I remembered you have a MKP10 in there and use MKP4. Looking carefully it looks like the MKP10 is C2 and on the first stage output. The MKP10 you have there is 630Vdc/400Vac, the...

If this works out I may put the design on a PCB or set - it should allow reduction of size and mess. I've designed and built two PCBs before - a 4 layer 24MHz clock board with low phase noise clock, 1:3 clock spanning IC and power filtering and shunt supply, a 4 layer 24MHz i2s isolator with...

Now running full B+ which means around 330Vdc B+ and 250Vdc on the plates. I've made a few changes: * running off mains and full voltages * heaters now running AC to keep voltages up * waveform is sampled from just above the plate but not directly at the moment. So it's more sensitive. *...

Mmmmm.... Currently setup on the bench supplies with a B+ of 61Vdc rather than 330-350Vdc. This allows me to test it and see it working whilst running on a current limiting power supply. I've then scaled the bias voltage down to 5V rather than 15V. So when you watch this - the input signal...

Ahh good choice on the Wima caps. I have FKP1, MKP10 and MKP4s here. I've also got Vishay axials in the turret board JCM. How much space do you have under those large power resistors?

Just a starting test to ensure this board works. The signal I'm using here is between 1 to 20V peak-to-peak, however in the compressor it will be subject up to 80V peak-to-peak You have to remember the tubes are cathode biased at 15V, so the grid is operating at -15V then this bias is layered...

It occurred to me last night, an option here issue the second ecc99 triode. I know TubeCad loves doing this for his headphone amps, but in this instance it works nicely. I may add a grid blocker to prevent RF interference on the output creating noise. This option provides a little more mA...

Been working through the resistances needed.. one is a 12K 8W cathode resistor for the output ec99 driver. Nope. So an option here is to use one of my DC SOA capable 950V MOSFETs.. Then run 3.5-4V on the gate and let the heatsink take the load. I think that will work. I think I won't have...

So the recycling of bits is going well... and yes this is a prototype hack fest! Next stop is to put some standoff tag boards, some pots and some jack holes.. I'm going to risk the 350V 390uF caps but I have some 450V 47uF if need be..

So the plan this weekend is simple: Stage 1 - check the variable volume works. 1. Make power supply by rip out the HV and heater power supplies from the old tube headphone amp. 2. Cut a wood board out for mounting hardware (I'll switch to alu once I know this step works) I'll basically use the...

Interesting. I don't think the Hz will always be accurate to two decimal places due to the 96Ksps max rate. The sample period time dictates the *accuracy* rather than simply the highest frequency. If you think about the more samples the more accurate you can measure the length of the frequency...

This is a real time build log.. so you're gonna get some LTspice first ... (mods if this isn't acceptable for the competition, let me know) I have a compressor. It's small, petite, is a Ross-style one which is.. ok.. but it's not great. My journey started with thinking - that Cali76 stacked...