That VFD driver normally has a 5V supply and a -36V supply to drive the VFD, with the filament being normally 3-6VAC, supplied from either a separate power winding on the transformer, or from a small isolated inverter block on the board, both with a centre tap. The reason Agilent went for the 18V supplies was likely because they likely did not want to have another winding on the transformer, to supply a -36V rail, or have an on front panel DC-DC converter unit to provide the -36V and the filament supply. That NEC microcontroller, with different mask ROM versions has been used in a lot of equipment, as it has enough IO pins to make it useful, plus also has the VFD drive and a few integrated peripherals on it. It has been used on many pieces of consumer equipment. The power transformer is common to many of the instruments, and they did not want to put extra windings there, that would be unused most of the time. The 18V rails are already there for the analogue stuff, so makes sense to use them to draw that extra current, and using both rails also ensures most of the current does not flow into the 0V rail, with only a tiny constant current from the driver being in the 0V rail. Level shifters for the clock and data into the display is then cheaper, and lower board space, as it is in otherwise unused volume under the display, while putting the DC-DC converter would limit the case volume behind the board, and in a 6.5 digit meter this would also need to have a really good magnetic and electrical shield (likely a fully soldered tin plated steel can) sticking right into the area above the sensitive input circuitry. Instead you just have the 2 steel shielding plates, connected via a RF ground to the chassis, to remove the display switching, and this works well.
Lovely meter, and great repair, as always. I loved the explanation of the (initially complicated, in my eyes) power supply to the microcontroller. Your detailed explanation of their rationale makes a lot of sense, of course, but I sure didn't think of it - which explains why I am better at blowing up stuff ;-) Have a very Merry Christmas and a happy, healthy New Year!
@@IanScottJohnston I have a bunch of these at work and all the displays are dim, probably been running for a couple decades. How to boost up the brightness of the display? On the HP E3633A (and all the similar style) supplies, the display is dim and what seems to fix it is the zener diode that offsets the filament voltage. Any comments on that for this DMM?
@@Ozzy3333333 i havent tried to fix this issue on these displays. Might be worth studying the schematic albeit my concern would be to blowing the VFD with any mods like that.
I may have done a video already, or at least included some detail of what I have been doing with GPIB in another video.......but no harm in doing a dedicated video. It's quite interesting I think.
It is found very useful explanation of repairing of hp 34001a dmm but kindly explain filament voltage, and the secondary ac voltage details of power transformer, waiting for your kind response, thanks
What are you going to do with the old VFD? Can it be repaired? Sometimes you can improve the brightness by increasing the filament voltage a bit. Its strange that only 1 segment is bad.
6:27 Last chance to take a photo of the original assembled unit so that on inserting the new display you can compare and be 100000000% sure that you don't have it rotated 180 degrees! (Or equally bad -- soldered to the wrong side of the board.) That kind of mistake is surprisingly easy to make, as one learns from painful experience! And prompts designers to make wrong-way insertion impossible by avoiding symmetrical pinouts etc.
@@IanScottJohnston I haven't worked on this same board. But I have seen myself and many others make this kind of mistake, and learn how easy it is to do, and take measures to avoid it that may seem ridiculous, but nonetheless catch such errors in the nick of time. Those errors happen perhaps because you didn't realize there was more than one way to fit the part. Or you think you can rely on some particular feature for orientation, but then you're working upside down, or one step removed from the orientation feature. Or the board has parts on both sides so there's no obvious "component side". Or mentally transforming a mark on one side to the other goes wrong . Or there's a phone call in the middle of your process. Or it's a part with unconventional pin numbering. In some way, your dependence on that one orienting feature, plus a slight error in observing it, causes you to carefully fit the part incorrectly. Usually noticed right after you finish soldering the thing as you admire your handiwork.
Yep.....but I have never done that, always found it good practise and builds up the skill level of removing large or leggy components. Our PCB's at work used to be single layer and the shock of cutting legs would/could lift pads from the PCB.
@@IanScottJohnston I found that with double sided PCB's getting the pin free from the plating in the hole was the hardest bit. With logic chips it was best to cut the legs as close to the body of the chip as possible and then just melting the solder and lifting the leg with the tip of the iron where it bent to enter the chip. We did use single sided boards back in the late 70’s when I first started, but soon moved over to double sided when we brought the 60’s designs up to date in to the 80’s. Plus we started to use CPU’s to control the audio switching in our avionic products around that time.
After watching the video, you did a great job getting it out without much trouble, if the replacement unit had the same issue as the presumed faulty unit - pointing to a fault elsewhere - you could reuse the vfd display.
![ALLY - OH MY! [ REACTION ] 'Amarin Nitibhon' #ALLY_OHMY #ALLY #allynitibhon #แอลลี่](http://i.ytimg.com/vi/nX9qpHjVfbQ/mqdefault.jpg)
Thanks for linking part 2 in the description
Glad you got a display. They are super stable and require next to no warmup. Happy Christmas.
I did order two from different sources......so I guess I'll have a spare when it arrives.
Good news + good display = good video. Merry Christmas. Cheers from Belgium
Did exactly the same job on my meter Ian & the Chinese sourced vfd is still good 5 years later
Great to hear!
Thanks!
I’m amazed you got a replacement VFD great repair. Merry Christmas and happy new year to you and the family. Regards Chris
I checked availability before buying the 34401A. There's pleny, at the moment, on Ebay. All the best to you and yours over the holibobs.
Good video
That VFD driver normally has a 5V supply and a -36V supply to drive the VFD, with the filament being normally 3-6VAC, supplied from either a separate power winding on the transformer, or from a small isolated inverter block on the board, both with a centre tap. The reason Agilent went for the 18V supplies was likely because they likely did not want to have another winding on the transformer, to supply a -36V rail, or have an on front panel DC-DC converter unit to provide the -36V and the filament supply. That NEC microcontroller, with different mask ROM versions has been used in a lot of equipment, as it has enough IO pins to make it useful, plus also has the VFD drive and a few integrated peripherals on it. It has been used on many pieces of consumer equipment.
The power transformer is common to many of the instruments, and they did not want to put extra windings there, that would be unused most of the time. The 18V rails are already there for the analogue stuff, so makes sense to use them to draw that extra current, and using both rails also ensures most of the current does not flow into the 0V rail, with only a tiny constant current from the driver being in the 0V rail. Level shifters for the clock and data into the display is then cheaper, and lower board space, as it is in otherwise unused volume under the display, while putting the DC-DC converter would limit the case volume behind the board, and in a 6.5 digit meter this would also need to have a really good magnetic and electrical shield (likely a fully soldered tin plated steel can) sticking right into the area above the sensitive input circuitry. Instead you just have the 2 steel shielding plates, connected via a RF ground to the chassis, to remove the display switching, and this works well.
Lovely meter, and great repair, as always. I loved the explanation of the (initially complicated, in my eyes) power supply to the microcontroller. Your detailed explanation of their rationale makes a lot of sense, of course, but I sure didn't think of it - which explains why I am better at blowing up stuff ;-)
Have a very Merry Christmas and a happy, healthy New Year!
Much appreciated....thanks!
@@IanScottJohnston I have a bunch of these at work and all the displays are dim, probably been running for a couple decades. How to boost up the brightness of the display? On the HP E3633A (and all the similar style) supplies, the display is dim and what seems to fix it is the zener diode that offsets the filament voltage. Any comments on that for this DMM?
@@Ozzy3333333 i havent tried to fix this issue on these displays. Might be worth studying the schematic albeit my concern would be to blowing the VFD with any mods like that.
Good news it's all fixed, maybe be a video of what your doing with the GPIB ?....cheers.
I may have done a video already, or at least included some detail of what I have been doing with GPIB in another video.......but no harm in doing a dedicated video. It's quite interesting I think.
Superb job Ian =D Great to see this meter returned to tip top condition! Have a great Christmas and New Year if I don't speak to you before then =D
was just wondering which seller you got this VFD from?
It is found very useful explanation of repairing of hp 34001a dmm but kindly explain filament voltage, and the secondary ac voltage details of power transformer, waiting for your kind response, thanks
What are you going to do with the old VFD? Can it be repaired? Sometimes you can improve the brightness by increasing the filament voltage a bit. Its strange that only 1 segment is bad.
Maybe you could build a clock and use it in that.
Segment D, i.e. the bottom row of all digits was knackered. No illumination at all, not sure it can be repaired. I'll inspect the pin though.
@@IanScottJohnston Probably a cracked weld inside the envelope for the drive to that row. not easily repaired
6:27 Last chance to take a photo of the original assembled unit so that on inserting the new display you can compare and be 100000000% sure that you don't have it rotated 180 degrees! (Or equally bad -- soldered to the wrong side of the board.) That kind of mistake is surprisingly easy to make, as one learns from painful experience! And prompts designers to make wrong-way insertion impossible by avoiding symmetrical pinouts etc.
You didn't do it.....did you?......oops!
@@IanScottJohnston I haven't worked on this same board. But I have seen myself and many others make this kind of mistake, and learn how easy it is to do, and take measures to avoid it that may seem ridiculous, but nonetheless catch such errors in the nick of time. Those errors happen perhaps because you didn't realize there was more than one way to fit the part. Or you think you can rely on some particular feature for orientation, but then you're working upside down, or one step removed from the orientation feature. Or the board has parts on both sides so there's no obvious "component side". Or mentally transforming a mark on one side to the other goes wrong . Or there's a phone call in the middle of your process. Or it's a part with unconventional pin numbering. In some way, your dependence on that one orienting feature, plus a slight error in observing it, causes you to carefully fit the part incorrectly. Usually noticed right after you finish soldering the thing as you admire your handiwork.
if the display is scrap then I would just the legs and remove the pins one by one. much easer to do and less prone to lifting tracks.
Yep.....but I have never done that, always found it good practise and builds up the skill level of removing large or leggy components. Our PCB's at work used to be single layer and the shock of cutting legs would/could lift pads from the PCB.
@@IanScottJohnston I found that with double sided PCB's getting the pin free from the plating in the hole was the hardest bit. With logic chips it was best to cut the legs as close to the body of the chip as possible and then just melting the solder and lifting the leg with the tip of the iron where it bent to enter the chip.
We did use single sided boards back in the late 70’s when I first started, but soon moved over to double sided when we brought the 60’s designs up to date in to the 80’s. Plus we started to use CPU’s to control the audio switching in our avionic products around that time.
It can definitely be hit and miss either way. I've also lifted pads where the desoldering gun has needed more effort/exposure to heat than expected.
@@sdgelectronics Merry Christmas and a happy new year
@@IanScottJohnston Merry Christmas and a happy new year
Personally, I just cut the legs then gently remove each led in turn. MUCH less chance of lifting a pad on an old PCB.
I would have snipped the pins on the old unit and removed the vfd, then removed each pin separately.
After watching the video, you did a great job getting it out without much trouble, if the replacement unit had the same issue as the presumed faulty unit - pointing to a fault elsewhere - you could reuse the vfd display.
wht is the point of that animation
Fun.........
mascot/channel identity