Hello,
I would like to ask for some advice on an issue which I am dealing with at the moment.
I am in the process of completing another DHT high voltage amplifier. This time is a PP with 4 DHT tubes running with an anode voltage of more than 1400v.
I am using toroidal PP transformers so equalizing the current is important. To achieve that I would need some effective way of monitoring the current the drawn by each tube at the same time. That could be accomplished with 4 panel meters with a 0-200mA range.
Also, unfortunately all tube are powered off the same filament transformer (12.6 CT).
To measure the current I have thought of a couple of methods:
1. measuring current flowing from the anode (maybe as a voltage drop across a 1ohm resistor in series with the anode) (but HV 1.5kv+ danger of insufficient insulation, etc).
2. Measure current directly through the cathode (from 0v to GND).
I'd go with n. 2 as it is the safest option really, but then we get to the real issue.
My setup would look something like this:
The meters are in the path from 0v to GND but my question is: will the meters give a correct readout? Could having them all connected throw them off?
Thanks
I would like to ask for some advice on an issue which I am dealing with at the moment.
I am in the process of completing another DHT high voltage amplifier. This time is a PP with 4 DHT tubes running with an anode voltage of more than 1400v.
I am using toroidal PP transformers so equalizing the current is important. To achieve that I would need some effective way of monitoring the current the drawn by each tube at the same time. That could be accomplished with 4 panel meters with a 0-200mA range.
Also, unfortunately all tube are powered off the same filament transformer (12.6 CT).
To measure the current I have thought of a couple of methods:
1. measuring current flowing from the anode (maybe as a voltage drop across a 1ohm resistor in series with the anode) (but HV 1.5kv+ danger of insufficient insulation, etc).
2. Measure current directly through the cathode (from 0v to GND).
I'd go with n. 2 as it is the safest option really, but then we get to the real issue.
My setup would look something like this:
An externally hosted image should be here but it was not working when we last tested it.
The meters are in the path from 0v to GND but my question is: will the meters give a correct readout? Could having them all connected throw them off?
Thanks
I would use matched valves.
Make sure your current drawn is the same for all of the valves.
Make sure your current drawn is the same for all of the valves.
Last edited:
Thanks for the suggestion bit I am afraid it is impossible... these are russian transmitting tubes which cannot really be matched in any reliable (or economically viable) way.
Inviato dal mio GT-I9301I utilizzando Tapatalk
Inviato dal mio GT-I9301I utilizzando Tapatalk
Look carefully at your schematic. All of the current meters are in parallel. The only way that circuit will work is if there are separate transformers, or separate windings on a common transformer for each output tube.
If I were doing this I would use a pair of Antek toroids each with two 12 volt windings.
If I were doing this I would use a pair of Antek toroids each with two 12 volt windings.
Unfortunately that is exactly the answer i feared. Thanks!
Inviato dal mio GT-I9301I utilizzando Tapatalk
Inviato dal mio GT-I9301I utilizzando Tapatalk
There are some fancy high side current sensing chips that can be used in the plate circuit, but would require a bit of solid state circuitry and most if not all of them are SMD. I used them in a prototype 650 volt power supply.
The basic concept is shown in Analog Devices application note AN105 on page "high side-4". I adapted this circuit to a TI INA168 chip and used it in the current limiting and metering circuit of a variable power supply that put out up to 750 volts at over 1 amp.
The app note is too big to post here, but Google will find it.
The basic concept is shown in Analog Devices application note AN105 on page "high side-4". I adapted this circuit to a TI INA168 chip and used it in the current limiting and metering circuit of a variable power supply that put out up to 750 volts at over 1 amp.
The app note is too big to post here, but Google will find it.
Maybe using 4 LEM units, with either both heater wires from each valve passing through the LEM, or just the anode?
For the heater wire option, the net AC mains current would be zero, but the net DC current would be the idle current in the valve. The LEM would need sufficient resolution at the idle DC current. Perhaps the CTSR 0.3-P would be appropriate.
For the heater wire option, the net AC mains current would be zero, but the net DC current would be the idle current in the valve. The LEM would need sufficient resolution at the idle DC current. Perhaps the CTSR 0.3-P would be appropriate.
Would measuring the voltage drop across a 1 ohm resistor from 0 to GND work?
Or maybe measuring the voltage drop across the same resistor but in the anode path?
the voltage differential between the volt meter input leads would be reduce to under 1v.
Or maybe measuring the voltage drop across the same resistor but in the anode path?
the voltage differential between the volt meter input leads would be reduce to under 1v.
But would the readings be reliable? The parallel nature of the circuit could interfere, could it not?
If one of the four sensing resistors has a lower resistance, the other tubes will draw current from that source. Sort of "crosstalk". Not sure if I managed to explain it correctly.... 🙂
Inviato dal mio GT-I9301I utilizzando Tapatalk
If one of the four sensing resistors has a lower resistance, the other tubes will draw current from that source. Sort of "crosstalk". Not sure if I managed to explain it correctly.... 🙂
Inviato dal mio GT-I9301I utilizzando Tapatalk
This means that 1ohm resistors in the anode path would be more reliable
Inviato dal mio GT-I9301I utilizzando Tapatalk
Inviato dal mio GT-I9301I utilizzando Tapatalk
For cathode measurement, Tubelab.Com said all.
For Anode, it is easier technically but less safer.
For Anode, it is easier technically but less safer.
Tubelab's suggestion involves changing the filament transformer. It is not something I can do unfortunately... but maybe I did not understand his advice.
Inviato dal mio GT-I9301I utilizzando Tapatalk
Inviato dal mio GT-I9301I utilizzando Tapatalk
Right. He suggests a seperate winding for each flament instead of a single DC suply for all hence isolating cathodes/flaments psyhically.
Another solution may be using a switch commutator with 4 positions connecting meter to rezistors in each cathode to the ground. This way you wont have to use another transformer.
From what I see, you're having four 1-ohm-resistors in parallel from 1 leg of filament to ground, any difference voltage you see on these resistors is due to their resistance tolerance, not because of different current in each tube.
Duong
I guess the only solution is measuring voltage drop across an anode resistor (1ohm of sufficient size)
Inviato dal mio GT-I9301I utilizzando Tapatalk
Inviato dal mio GT-I9301I utilizzando Tapatalk
Did you not want to use LEM types sensors, or perhaps you didn't know how they work, or ?
You can use a hall effect current sensor like the acs712. The only problem you have it is required the use of micro to do the convertion. Just open the circuit for each filament and wired the sensor in series. Then program the micro to read each sensors and do the convertion. It looked simple but it is simple when you used a micro.
The acs712 is a little frail on working voltage, as the anode is the only single wire accessible point in the OP's amp for each valve to make a DC bias measurement.
Hi,
similar to tubelab´s suggestion I´d insert a measurement shunt Resistor between the heater and gnd and use a currrent sensing device like the INA138/168 or the INA250 (w. integrated shunt resistor).
The INAs function best with voltage drops of 50-100mV dropping over the shunt resistor, which is low enough to have enough voltage for the heater filament.
jauu
Calvin
similar to tubelab´s suggestion I´d insert a measurement shunt Resistor between the heater and gnd and use a currrent sensing device like the INA138/168 or the INA250 (w. integrated shunt resistor).
The INAs function best with voltage drops of 50-100mV dropping over the shunt resistor, which is low enough to have enough voltage for the heater filament.
jauu
Calvin
- Status
- Not open for further replies.