our new product - 5V temperature compensated voltage reference

[fcb]

I'm putting this in Metrology as there might be a few questions from the usual suspects.

Just released our new 5V 0.01% class precision voltage reference, this unit has an active temperature compensation system and is based on the REF50xx IC from TI (which has an on-die temperature sensor).  We will release the PC calibration software on our website in due course, as well as the protocol if someone wants to engineer there own calibration enviroment.  We are shipping in 2-3 working days as we carry out a final calibration.  We are specifiying it at 0.01% (100ppm) at the moment, although the results we are getting from early units are considerably better.

It's on eBay as well as our website.

http://www.ebay.co.uk/itm/124699468047

[szszoke]

Any chance this will be available from a supplier within the EU?

[alm]

Is 100ppm just initial accuracy, or is it guaranteed for some period of time? Do you have any data on the residual temperature coefficient (after compensation), noise or long-term stability? Or is this all within the 100ppm spec?

[fcb]

The 100ppm is for the first year.

The major component is aging. The TC system keeps the output typically within 5.00002V and 4.99998V (+/-4ppm) over the calibration range. Waving hands over it causes bigger short term effects.

[fcb]

Any chance this will be available from a supplier within the EU?

Sorry - at this stage we don’t have a distribution network in the EU. So are supplying from UK.

[peter-h]

A nice product especially for the money

Have you thought about doing a 2.5V version? It would be handy for calibrating the very common 2.5V references.

No actual need for distributors in the EU; one can sell direct. Germany and France are "culturally difficult" (businessmen there will buy foreign but look for local language support) and more likely to need local agents but if you have a good product at a good price...

[alm]

No actual need for distributors in the EU; one can sell direct. Germany and France are "culturally difficult" (businessmen there will buy foreign but look for local language support) and more likely to need local agents but if you have a good product at a good price...

This used to be the case until the results of the Brexit took effect. Since then, for me as an EU citizen, getting products from the UK takes substantially longer and is much more expensive then before. For me at least, buying from the US is less hassle and for light items competitive on price to buying from the UK. I'll think twice before ordering something from the UK again.

[Stray Electron]

   Alm, this is not a criticism but it would help if you posted your location so we'd know what part of the world that you're talking about.

[peter-h]

"until the results of the Brexit took effect. Since then, for me as an EU citizen, getting products from the UK takes substantially longer and is much more expensive "

You are dealing with stupid companies.

I've been in business for 40+ years. International trade the whole time. Currently there are some airmail issues (delays in Customs) in some countries, notably Hungary and such, but most of it works fine. The problem is probably that some UK vendors are too stupid and too lazy to generate the right paperwork. But the same goes for firms on the mainland; recently I had to buy some bike parts from France, via a friend there, because the stupid vendor could not send an airmail package to the UK with an invoice There has always been a % of companies who are too dumb to generate even an invoice. This is the same problem everywhere, not just the UK. Royal Mail's recent website changes have been an extra challenge; Royal Mail is as dumb as any other national postal service and they struggle with even simple website technology

[Ice-Tea]

Doesn't have all that much to do with being too dumb to generate an invoice. Shipping is a *lot* more expensive if customs are involved and the likelyhood you'll need to follow up and invest additional time and money to get it moving again is exponentially higher.

On a 27£ part it makes little sense.

[peter-h]

"Shipping is a *lot* more expensive if customs are involved"

It isn't. It doesn't cost a penny extra, and you have to generate an invoice anyway for your accounting requirements, so may as well produce it in the right format.

Packages always used to get lost. It's a cost you build in and is usually a lot cheaper to send stuff "untracked" and accept some % of loss or fraud.

[Ice-Tea]

"Shipping is a *lot* more expensive if customs are involved"

It isn't. It doesn't cost a penny extra, and you have to generate an invoice anyway for your accounting requirements, so may as well produce it in the right format.

Packages always used to get lost. It's a cost you build in and is usually a lot cheaper to send stuff "untracked" and accept some % of loss or fraud.

Oddly, my DPD invoices don't agree.

[peter-h]

Ah, DPD..... they hate doing anything which involves any e-f-f-o-r-t :smile: They are cheap though.

Then there is Yodel, who are even cheaper and who hate "effort" even more I had a delivery by them attempted to my office 25 (25) times. Each time the idiot swiped the barcode at a local petrol station instead and reported there was nobody at the office.

DPD have said they don't want overseas business at all.

I would try Royal Mail Tracked and Signed. Probably slower but generally works.

[fcb]

We are doing 5.000V, 4.096V, 3.000V, 2.500V and 2.048V, and have stock of each, although most of production was the 5.000V version.  We'll list the other variants soon on the website.

The situation with EU sales - we are happy to ship to anywhere in the world! Also happy to talk with anyone who wants to distribute our products in the EU - BREXIT is not making things easier would be an understatement.

With a low-cost product like the REF50X the extra import duty management fees are a disproportionate extra cost. Not alot we can do about that, and we wouldn't be able to factor in a distributor margin into the REF50X price to support a dealer (so they would have our blessing to sell it at EUR27+their margin) - likely end up costing the same as if we ship it direct...  It's on eBay and our website, and we make a little bit more if you buy direct but we are happy to sell via both.

The REF50X calibration is fairly horrific as it is semi-automated at the moment and represents a good chunk of the cost. We've found calibrating at 8C through to 32C, with upto 10 points seems to keep it within a +/- 4ppm band.  Our acceptance criteria for release is +/-10ppm.  Pictures from the EPIC REF50X sw during calibration.

Here's some of how it works - the processor is a PIC12F1572, and the code was written in assembler.  There are 128 locations at the top of program memory that have HEF (high-endurance flash), these are treated as 64x 16 bit memories.  The PIC ADC is 10 bit and is fed from a differential amplifier that amplifies/offsets the reference IC temperature sensor - some maths takes these values, uses the table (64 entries) and calulates the required PWM.

PWM is 1200Hz, which was a chosen as it's a multiple of 50 and 60Hz.  The PWM drives an analog switch that toggles between 0V and a low-pass filtered output from the reference IC, the output of the analog switch is then low pass filtered, buffered and low pass filtered again before being fed to the TRIM on the reference.  The range of adjustment on the TRIM can be quite extreme, from 4.9 to 5.3V, so the amount of movement in the PWM is pretty small over the compensated range.

Coating the reference IC in silicone conformal coating reduces humidity effects somewhat and we age the reference for at least 4 days at an elevated temperature. We'd need to sell a few thousand of these to pay back development costs, but the project has been a useful path-finder for potential future products.

We will publish the communications protocol at somepoint, and the sw is included in the next update to EPIC (Electron Plus Instrument Control) - it's under Instrument>Change Instrument on 21.007 onwards.  The 6 pin-header (TX/RX/LED/+V/0V/0V) will probably grow to 10 pins on the next revision for various reasons.

*Just noticed pix rotated - pix from my phone the other day..

[cemelec]

Just mention the word "Brexit" seems to inflame some people to make patently absurd statements. Over the years I have bought loads of stuff from China - incidentally never in the EU - without any customs or other problem whatsoever. (Banggood, AliExpress)

I have absolutely no doubt anyone can buy from the UK or EU without any problems.

[Ice-Tea]

1) You may want to glance over this:

https://www.reddit.com/r/Aliexpress/comments/eg36q1/why_sellers_can_ship_things_so_cheaply/

2) It's not because Chinese companies can shrug off accounting rules that everybody else can.

[fcb]

Just mention the word "Brexit" seems to inflame some people to make patently absurd statements. Over the years I have bought loads of stuff from China - incidentally never in the EU - without any customs or other problem whatsoever. (Banggood, AliExpress)

I have absolutely no doubt anyone can buy from the UK or EU without any problems.

Not quite sure what you mean.  BREXIT has and is causing issues with UK exports to EU. Not insurmountable but there are some real issues, the vast majority of the stuff we've shipped to the EU has been fine, the odd parcel has had problems that do get resolved (we recently had a parcel stuck in UPS Stanford Le Hope depot since 19 April, we just sent out a replacement via another courier).

There are some big changes happening from July for EU to EU transactions (doesn't obviously effect UK) - so fingers crossed it will all settle down later this year.

[cemelec]

Agreed, but it's not shipping costs I was referring to. I pay them from China, it's nice that usually they are quite low, but I would still buy if shipping were considerably more.

I remember a time before the EU single market existed, (I'm no youngster) and there were no problems then, so I don't see why there should be a problem now, other than political.

[fcb]

My thought is that if BREXIT was allowed to proceed smoothly and without pain or disadvantage (the "have our cake and eat it" scenario) then there might be a few EU countries keen to follow us out.

I've only everknown the UK in the EU and I really like(d) spending time in the EU, so pretty was pretty sad in 2016.

[peter-h]

I would be very interested in buying one of the 2.5V references, when it becomes available.

[fcb]

I would be very interested in buying one of the 2.5V references, when it becomes available.

Putting the other voltages up on the website this week, along with some other stuff.

[mendip_discovery]

I have bought one and I have a little bit of feedback for the Certificate,
I like there is a Cert Number, Lab Conditions, but it would have been nice to know the actual measured readings where so that I can record that in a spreadsheet, plus no measurement uncertainty.

Another query, when doing the test are you only heating the 5V source and not the whole board including the wires?

[fcb]

Hi mendip_discovery,

It's a balancing act, the longer term goal with the REF50X and other VREF style products is to have an auto-documenting calibration system that will include every measured parameter and generate a more comprehensive calibration cert (error bars, etc..).

How we calibrate the REF50X: Place it in a chamber (Design Enviromental 125L), this is taken down to 8^oC then brought up to 32^oC over several hours.  EPIC (our SW) has a REF50X interface and we calibrate tweak the DAC settings for 5.0000V every few ^oC. Lightly twisted banana leads connect the REF50X to the DMM on the outside of the chamber (does this answer your q?).  We don't record the actual temperatures that we take each point at.

Once we have enough points, EPIC calculates the the missing points and uploads a table of 62 ADC<>PWM offsets into the RFE50X (values 0 and 63 are trigger values for under and over temperature LED display). The SW in the REF50X performs an interpolation between two points to get a finer degree of adjustment on the PWM DAC.

We have often see +/-2ppm over the 10 to 30 range short-term, but this is getting to edge of what we can do at the moment.

EPIC 21 (currently 21.008) on our website can connect to REF50X if you want to poke about.

[peter-h]

Great product. Bought a 5V and a 2.5V one.



Funny, testing them against a 30 year old HP multimeter



HP knew their stuff in those days!

An ancient 3468A is also within 0.01%.

[mendip_discovery]

I bought one of the 5V ±10ppm references (s/n 8R3S1B) and I thought I would test it on my 121GW and that reads some 5.0011V and -5.0004V which was confusing me.

So I took it to work and tested it on my trusty Agilent @20°C I got a reading of 5.00050V and -5.00051V, taking it further I put the Agilent onto the Transmille and got 5.00000V both ways. I did this during my lunch break but so far I would say its about 100ppm not 10ppm out.

I dug out the cert that I got with the 5V reference and it irritatingly gives me a measured value of 4.99995 to 5.00005 and no Unc so I can't check my EN ratio so if one of us has an issue. Though I am not worried about the work kit, did a check with UKAS a month ago and they were happy I was well within my Unc.

btw don't see this as me bitching about a product, just reporting my issues. Though I must admit I would like to test it another day to be sure.