I was tinkering the other day for a simple solution for a project with a DC filament/heater power supply.
These days I often like to use these small DC SMPS power supply (from Meanwell for example).
The downside, is that they are a little less nice on the heaters because of the inrush current.
So I just made a super simple PMOS soft start.
Since I had a simple capacitance multiplier in the HT/B+ voltage anyway, I figured that it would be nice to connect the two together.
It also detects when the heater voltage drops.
This is pretty crude initial drawing, most values are not critical, except the resistors for the soft start, which can be changed according taste.
D1 and D2 are maybe a little overkill, I didn't test this in practice yet, so maybe to can be left out or replaced by a resistor.
The value of the zener diode needs to be close to the heater voltage.
R12 is just there to release some heavy lifting on Q2.
Otherwise there will be a huge but very small power spike when being switched.
R9 is just a test load, can be dismissed.
I would recommend putting a small 22uF or 47uF capacitor on the output of M2.
If people are interested, I can show some switching plots, but there isn't anything fancy going on or to see.
The soft start delay is now roughly 2 seconds, after that M2 will switch on to be at full about 0.5-1s later.
I am open to some simplifications or other ideas.
The goal here was to be cheap and simple, and I think this will just work fine in practice 🙂 🙂
These days I often like to use these small DC SMPS power supply (from Meanwell for example).
The downside, is that they are a little less nice on the heaters because of the inrush current.
So I just made a super simple PMOS soft start.
Since I had a simple capacitance multiplier in the HT/B+ voltage anyway, I figured that it would be nice to connect the two together.
It also detects when the heater voltage drops.
This is pretty crude initial drawing, most values are not critical, except the resistors for the soft start, which can be changed according taste.
D1 and D2 are maybe a little overkill, I didn't test this in practice yet, so maybe to can be left out or replaced by a resistor.
The value of the zener diode needs to be close to the heater voltage.
R12 is just there to release some heavy lifting on Q2.
Otherwise there will be a huge but very small power spike when being switched.
R9 is just a test load, can be dismissed.
I would recommend putting a small 22uF or 47uF capacitor on the output of M2.
If people are interested, I can show some switching plots, but there isn't anything fancy going on or to see.
The soft start delay is now roughly 2 seconds, after that M2 will switch on to be at full about 0.5-1s later.
I am open to some simplifications or other ideas.
The goal here was to be cheap and simple, and I think this will just work fine in practice 🙂 🙂
The problem is given in the description, current in-rush on the heaters when power supply is switched on.
Also you really want to make sure to turn on the HT/B+ after the heaters have been turned on.
Which is being done automatically in this case.
I don't call that a non-existing problem?
I call a handful of passive and some very standard discrete components that cost close to nothing, simple yes.
They are cheaper than even a 3 position switch.
The fets were already there to begin with.
It has been done now for around 100 years, applying B+ and heater voltage at the same time. With cool heaters there will be no current flow (in vacuum). Switching on the heaters wil slowly allow current flow between Cathode and Anode. Heaters (at least most of them) were designed to apply the heater voltage at once.
I know a delayed heater voltage can be made, but I see no reason to do so. On the other hand I do use soft start electronics for power transformers (> 500 Watt). That helps to limit the current surge on startup.
Perhaps a switching power supply cannot drive the (extra) heavy load of a cold heater, but that’s another story.
Regards, Gerrit
I know a delayed heater voltage can be made, but I see no reason to do so. On the other hand I do use soft start electronics for power transformers (> 500 Watt). That helps to limit the current surge on startup.
Perhaps a switching power supply cannot drive the (extra) heavy load of a cold heater, but that’s another story.
Regards, Gerrit
Some switching power supplies have a tendency to trip with the initial current spike.
Especially when there are quite some additional output tubes.
Which is/was the main purpose of the soft start circuit. (which I already mentioned in the first post)
There are many things done for 100 years, doesn't mean they all work optimal.
In my experience some tubes just last longer/better when you just don't hit the heater with a current in-rush on a cold start.
I have actually blown a couple that way.
It's just a nicer way of doing things, don't see what's wrong with that?
That's like 99% of circuits and topics here anyway?
In this case it just a less than a dollar, and will automatically detect when heater voltage all of a sudden disappears.
Like I said, the FETs are there already anyway for the high voltage and soft starting against inrush.