How to kill a iN4007

[ruffrecords]

I have been playing around with this HT SMPS PCB and got it basically going (outputting 250V square wave at 50KHz). It converts 12V dc to tube HT type voltage.

The I added a diode (UF4007 I thought) and a 100uF/400V cap. No problems. Well nearly 400V dc on the cap. So I tried adding about 35mA of load resistance. Seemed OK initially but the second time I turned it on my bench supply went into current limit and the and its output volts dropped to 7.5V. At first I thought I had blown the cap but it still would not work with just s 22uF/400V cap. So I replaced the diode and it worked. SO I decided to add just a few mA of laod and the same thing happened. Checked the diode and it was shorted.

Now I thought these were UF4007 types which should have been perfectly OK. Turns out I had some old 1N4007 left in the same drawer and that is what I had been using. Proof positive of the need for fast diodes in a SMPS.

Cheers

Ian

 

[trobbins]

Yeh some parts are worth buying in batches, and quarantining the batch (ie. keep them in any plastic packet that came in), even for DIYers.  Much easier then to have confidence that a part in the drawer is new and from the same batch, or just a loose item thrown in the drawer in days of yore.

Parts from the same batch can be very useful for special applications (eg. series connection of diodes for higher PIV).

 

[JohnRoberts]

I have been playing around with this HT SMPS PCB and got it basically going (outputting 250V square wave at 50KHz). It converts 12V dc to tube HT type voltage.

The I added a diode (UF4007 I thought) and a 100uF/400V cap. No problems. Well nearly 400V dc on the cap. So I tried adding about 35mA of load resistance. Seemed OK initially but the second time I turned it on my bench supply went into current limit and the and its output volts dropped to 7.5V. At first I thought I had blown the cap but it still would not work with just s 22uF/400V cap. So I replaced the diode and it worked. SO I decided to add just a few mA of laod and the same thing happened. Checked the diode and it was shorted.

Now I thought these were UF4007 types which should have been perfectly OK. Turns out I had some old 1N4007 left in the same drawer and that is what I had been using. Proof positive of the need for fast diodes in a SMPS.

Cheers

Ian

How hot does the diode get? Shorted sounds like it literally melted inside.

I once made a very HF cap doubler/tripler using the square wave from a switching power supply with a couple 0.1uF caps and small switching diodes (1n4148 or equivalent) to generate a few mA @ 48V for phantom power.

JR 

 

[ruffrecords]

How hot does the diode get? Shorted sounds like it literally melted inside.
odes (1n4148 or equivalent) to generate a few mA @ 48V for phantom power.

JR

I do not know. I tall happens very quickly. The diodes did not feel hot when I removed them.

Cheers

Ian

 

[PRR]

> very HF cap doubler/tripler using the square wave from a switching power supply ... and small switching diodes (1n4148)

'4148 is 4,000 times faster than '4007.

1N4148: Reverse recovery time IF = IR = 10 mA, iR = 1 mA    trr 8 ns
1N4007 has a typical reverse recovery time of 30 μs (note: u not n !)
ultrafast version UF4007 has Trr = 75 ns

Ian just grabbed the wrong part.

 

[ruffrecords]

Paul is spot on as usual. UF4007 arrived in post this morning. Successfully connected it to 100uF 400V cap plus 35mA load and ripple monitoring network. Result, no dead diodes, close to 225V dc and very short bursts of 50mV pp ripple every 8uS or so.

Output power is about 225 x 35 / 1000 = 7.875 watts

Input power is 12V at 0.75A = 9 watts.

So efficiency is 100 x 7.875/9 = 87.5%

Not a bad start.

Cheers

ian