HP85644A voltage selector?

[G0HZU_JMR]

I've bought an old HP85644A tracking generator source on UK ebay as spares/repair and it was described as "does not power up"

http://www.ebay.co.uk/itm/Hewlett-Pa...vip=true&rt=nc

It's arrived this evening and it appears to have the fault that I dreaded... the input voltage selector appears to be set to 115V

The main on/off switch on the front panel is obviously toasted because it won't 'click' anymore. There is a very faint smell of burnt bits if I sniff through the fan but I might be imagining it...

I haven't tried switching it on yet but I'd like to ask for advice on what usually happens (on HP test gear in general) if the selector switch is wrongly swapped over and if there are any clever checks I can do before trying to power it up. I'm guessing the fuse rating will be higher for 115V so there could be some serious carnage inside if someone in the UK has powered it from 230V

Worst of all the service manual isn't very comprehensive so I don't even know yet what type of PSU there is inside. I kind of hoped HP would design it such that it would survive this type if misuse but maybe the previous person kept trying it again and again until the switch failed?

Any advice on mains inlet checks most welcome because this is one area in electronics where I lack decent experience.

[WME_bill]

I fear I will have to say Bad Luck.
I did that once with a bit of Admiralty equipment, set at 115v and it burnt out the mains transformer. If you are lucky, the smell may be only the fan, which will be 115v, but easy to replace. HP tended to keep to linear power supplies, rather than switching. If it does have a switched mode supply, then you are going to have great fun replacing the switching transistors and control circuitry.
Either way. do not plug it in.
You will have to open it up, and see and smell to find what has been burnt.
I will see if I have a circuit or manual, but not very hopeful. wme_bill.

[G0HZU_JMR]

Hi Bill
I've been skimming through the rather brief service manual and found some text about the input protection. See the linked image below.

It does look like there's a design attempt to provide protection against this issue but it also states that if the fuse doesn't prevent the 230V getting past the fuse then severe damage will result to the switching PSU.

Oh dear...

Because the front switch is obviously damaged I fear that the previous owner tried several fuses, several times. The only faint hope is that there isn't an obvious smell from the unit.

Also, I need to be sure it isn't 'me' that's misreading the selector switch. I assume that if you look inside the hole and can see 115V then it's set to 115V.

There is a recessed red plastic slider switch that can be operated by a pen/screwdriver and it shows 115V when I look down the hole. If I slide the switch to the other position the 115V logo disappears and it then shows 230V. So it looks like it was set wrong. It might work now but I don't think it's wise to try...

[G0HZU_JMR]

Good news... very good news...

I took out the fuse carrier and looked at the fuse. The fuse carrier sits in a little pullout plastic shelf under the mains inlet and it can be levered out with a screwdriver. I looked very carefully at the holder and there were no marks on it from previous 'screwdrivers'

Inside the holder is the inline fuse (blown) and also a SPARE FUSE! Nice one, HP !

They were both the same fuse type and rating and the spare was healthy! So I put the spare fuse into the holder and swapped the selector switch to show a visible 230V logo and plugged it in....


It seems to be working!

It all lights up and the display does its bootup sequence OK.

Obviously I'm going to have to replace the front panel mains switch because it is stuck 'on' and won't click off but it looks like it's going to be a bargain after all!

It can be used as a 6GHz sig gen as well as a tracking source so I will probably use it for this rather than as a tracking gen for my HP8566B spectrum analyser

The image below shows the blown fuse and the spare one now mounted in the carrier ready to try.

[ronbryan]

Jeremy

Could the smell be coming from the IEC mains input filter? I know the ones supplied by Schaffner are very unreliable after about 10 years as the filter capacitors fail - I have had a couple of HP comms analysers fail in this way.

https://www.vintage-radio.net/forum/...7&postcount=25

Ron

[Radio Wrangler]

Ah! Good.

There was a test in HP's formal design verification process to test that powering up on the wrong range would pull the line fuse (provided the right fuse was fitted)

In old PSUs with iron transformers, it used a carefully designed transformer which would saturate and pull the fuse, with a saturated 50/60Hz fan helping sometimes.

Later SMPS types had added protection stuff. Sometimes a crowbar, sometimes big zeners. They were only specified to work once. After a fuse has been pulled it was assumed someone would give the unit a check-over and fresh zeners. In some units we used Siemens gas arrestors which could live through several fuses. The slide switch for two range supplies strongly suggests a switcher, heavy iron units had at least 4 ranges. The slide switch converts the bridge rectifier to a voltage doubler.


David Never been inside one of those. I'll need to look for one for my 8566B

[G0HZU_JMR]

Hi Ron
I'm not sure but the very faint smell from it is probably just 'dusty electronics' smell. It doesn't have the nasty edge to it when a cap or semiconductor fails.

It could be from the front panel switch contacts though? Maybe the current surge heat welded the switch in one position as the contacts touched when the original owner blew the fuse?

Never been inside one of those. I'll need to look for one for my 8566B

Hi David. Here's an image of it hooked up to the analyser in cw mode at 5GHz. It isn't fully synthesised when in standalone cw mode and it's very slightly off frequency (but probably still in spec).

The RF level range (via internal attenuators?) is from about +10dBm down to -80dBm and the attenuators check out OK right down to the -80dBm limit.

I'll try using it in tracking gen mode with the 8566B over the weekend. But I think I'll need to consult the manual for that as I need to couple across the LO signal and probably the 10MHz ref as well.

But it's looking a lot better than it was when I started the thread a few hours ago

[Radio Wrangler]

Nice!

David

[G0HZU_JMR]

These tracking sources date back to about 1990 and are quite rare but I managed to find a user manual online today and it showed me how to connect it up as a tracking gen for the HP8566B.

I didn't realise it supported so many other instruments but you have to select HP8566 from the internal menu system. Navigating the menu system is a bit awkward because the display only has two lines of text but I managed to get it set up OK for the 8566. There are several BNC cable connections to make between the two instruments as well.

Initial checks are very impressive in terms of tracking accuracy and system flatness. I tried it on a 0 to 2.5GHz span and also 2.5GHz to 5.8GHz. See the plots below on 1dB/div. I didn't try normalising the traces so these responses are just the raw uncorrected response. Very impressive indeed!

Obviously, the 'one to have' is the 26GHz version (HP85645A) of this tracking source but these are quite rare and still command huge prices. I think the 26GHz versions were about $30k when new and the 6.5GHz versions were about $20k new. I got this one for £176 + P&P which I think is a bit of a bargain even though its well over 20yrs old

I'm not too keen to leave this one switched on unattended until I've had a good look and sniff inside as Bill suggests. I'll also have a good look around the mains inlet for signs of damage/overheating but I'm hoping the only casualty is the mains rocker switch. It's probably unsurprising that it has welded its contacts as the overvoltage protection is probably designed to blow the fuse very quickly and reliably. So there must have been a huge surge current when the rocker contacts first touched together... . Hopefully the mains rocker switch is cheapish and easy to buy and replace.

I'm guessing this instrument was imported from the US and somebody in the UK switched it on without swapping the voltage selector over. It must have made quite an angry bang when it blew the fuse! I then suspect the owner didn't know about the mains fuse hidden inside the mains inlet and assumed the unit was toasted because of the obvious damage to the mains on/off rocker switch and the loud bang it must have made. So it ended up on UK ebay for spares/repair?

The plots below are on 1dB/div and are not normalised. I used a decent HP branded N type cable to connect the two instruments together. The response is remarkably flat!

[Radio Wrangler]

The 26GHz version has to contain a YIG oscillator, and the harmonically tuned filter from the 8566B in order to generate the higher bands. Essentially it's the RF section of the analyser running backwards. They don't 'arf get hot. Maybe the 6 gig little brother will be more reliable?

David (Just ever so faintly green-tinged)