help with: ZD850 1.5 Amp Constant Current Driver

[imageicon]

Hey guys i am looking at building some more bikelights and to cut down costs i am looking at cheap constant current regulation and i found this piece of tech.
http://www.cutter.com.au/proddetail.php?prod=cut755

as i understand it its just to hook up power leds and a resistor between RLIM and PGND ... it says in the datasheet that regulation is determined by an equation 0,2V/(1.15*iLED) where iLED being current wanted so say i want 1000 mah .. can any of you tell me wich resistor to put in?

what else needs to be done i see the typical application diagram and it seems to require some more work but is that necessary?

Cheers Jesper

 

[frenzee]

For 1A, you'll need a .174 Ohm resistor. You could make your own from a 5 3/4" length of 24 ga. copper wire. btw, I think your link is for the chip only. You'' still need to put together the rest of the board or get the eval board.

 

[imageicon]

ya it dosnt say if its the eval board or the chip.. id like the eval board i think... that was the resistance i calculated too it just seemed so small that i thought i was wrong!

think ill email them ask if its eval board or what it costs!

 

[macforsale]

*

 

[imageicon]

hey guys i wrote to mark " cutters ... the eval board is 10$ not to bad for a driver ... so do you guys recon thats a full functional led drive or do i still need to work on the circuit? im abit of a noob with elektronics and the whole led thing...

 

[Luminescent]

hey guys i wrote to mark " cutters ... the eval board is 10$ not to bad for a driver ... so do you guys recon thats a full functional led drive or do i still need to work on the circuit? im abit of a noob with elektronics and the whole led thing...

Click on the link at cutters and download the eval board manual (here's the link):

AN020_ZD850EVB_MANUAL_07 30 07.pdf

Here is a quote from the beginning of the manual.

ZD850EVB Evaluation Board Factory Default Settings
The ZD850EVB was configured to operate at a voltage supply
input of 12VAC/VDC to the VAC1/VAC2 of the device
driving 3 high power LEDs in series with ILED current of
300mA~350mA (details shown in Figure 4 under the Typical
Applications Section).

For 350 mA operation they show a 220uH inductor being required, and it looks like this MIGHT be what is supplied on the board based on the pictures ('Zywyn' didn't bother to SHOW the schematic and parts list of their eval board right up front like EVERY OTHER MANUFACTURE ON THE PLANET, so it's a little hard to tell). There is a 'Bill of Materials' at the end of the manual, but it suspiciously fails to list the inductor, so it's anyone's guess.

So 12volts supplying 3 White LED's at 350mA is pretty much guaranteed right out of the box, but higher current will require a different 'current sense' resistor, and possibly a new inductor.

The board seems to be mainly programmed by changing a single current sense resistor, but there are limitations based on the desired current, the number of LED's in series, AND the supply voltage.

For example for a 12-16 volt supply with 3 LED's and 1.5 amps of current, figure 5.3 shows that the current sense resistor will drop down to 0.06 ohms, but it also shows a different 68uH inductor.

The resistor value seems to be a little different from what the formula predicts so I guess there is some fudge factor based on actual testing needed.

So it's not quite as simple as you thought, but you should be able to deal with it if you can deal with fabricating a very low value resistor and bridging it into the PCB. It doesn't look like they provided terminals for this so you would have to bridge across an existing TINY surface mount resistor.

The inductor issue is a major pain in the *** because if you set the current too high for the inductor supplied, the core will saturate and the current will suddenly go very high and most likely destroy either the ZD850 chip, or your LED's (or BOTH).

Unfortunately, as I mentioned above, the inductor that is supplied doesn't seem to be spec'ed anywhere in the ZD850 Eval Board manual.

Without this info, and a little more information about just WHAT YOU PLAN TO DO it's a little hard to tell if this will work for you or not.

Can't do much about the missing Zywyn info execpt contact the company, but perhaps we CAN nail down what you want to do a little more accuratly.

- How many, and what type of LED's do you want to drive?

- What current were you planning to drive them at? (you mention 1 Amp).

- What battery voltage [or voltage range] are you planning to use?

If you are planning to use NiMH batteries to drive the LED's, they have a nice flat discharge curve, so you might be able to find a nice simple linear regulator constant current circuit and still get 80% efficiency or so.

 

[macforsale]

*

 

[Luminescent]

Luminescent,
Why trash the manufacturer? Nothing out of line with their data sheet.

I ask that you edit your post. The forum's membership is not well represented by your comments.

mike

Ok, in deference to your sensitivities, I have toned it down.

The data sheet is indeed reasonable, but the "evaluation board documentation", is just a BAD rehashing of the datasheet, the included schematics are at CHIP level NOT at BOARD level, so it is nearly USELESS for determining how to wire up the eval board.

Call me silly, but I come from the old engineering school when 'evaluation board documentation' at least told you SOMETHING specific about the eval board.

The closest Zywyn comes to actually explaining anything specific about the eval board, is to say that details can be found in Figure 4 (except figure 4 is really 4.3 and is just a general chip level schematic that doesn't give ANY specifics like the inductor part number or current ratings, or diode part number or voltage or current ratings, or pcb layout or wiring diagram)

True, I have seen a lot worse, but the fact that the eval board docs lacks even a simple wiring diagram or hookup information for the eval PCB alone makes it totally unacceptable.

 

[WeLight]

Ok, in deference to your sensitivities, I have toned it down.

The data sheet is indeed reasonable, but the "evaluation board documentation", is just a BAD rehashing of the datasheet, the included schematics are at CHIP level NOT at BOARD level, so it is nearly USELESS for determining how to wire up the eval board.

Call me silly, but I come from the old engineering school when 'evaluation board documentation' at least told you SOMETHING specific about the eval board.

The closest Zywyn comes to actually explaining anything specific about the eval board, is to say that details can be found in Figure 4 (except figure 4 is really 4.3 and is just a general chip level schematic that doesn't give ANY specifics like the inductor part number or current ratings, or diode part number or voltage or current ratings, or pcb layout or wiring diagram)

True, I have seen a lot worse, but the fact that the eval board docs lacks even a simple wiring diagram or hookup information for the eval PCB alone makes it totally unacceptable.

These are not unreasonable comments, I will forward to Zywyn to try to get a little more 'meat' on the detail

Cheers
Mark

 

[WeLight]

I have just uploaded a newer Eval board datasheet which addresses these questions
http://www.cutter.com.au/proddetail.php?prod=cut755

Cheers
Mark

 

[macforsale]

*

 

[Luminescent]

Gets one to the same document that 'Luminescent' used, but he made all the comments about missing information. :shrug:

Oh it's different all right.

Actually, it IS a differently dated version but, sadly, actually seems to be WORSE in some respects :sigh:

For example the figure I thought was called out as 4 but which looked like 4.3 due to funky formatting, actually WAS "Figure 4. _ 3 LED something or other" but was jammed together by poor formatting which made it look instead like "Figure 4.3_LED something or other"

Now they still call out the 3 LED example at the beginning as "Figure 4" but the actual 3 LED example in the main body of the document has been change and is now "Figure 5"! (So it went from slightly confusing to flat wrong).

Other than the above NEW error in referencing the proper figure number, their doesn't seem to be any change at all. :shrug:

It was damn nice of Mark to make an effort to respond so quickly, but the document is now apparently worse than ever, and still seems to lack even a minimal "Hook-Up Diagram" at *BOARD* level, and full parts breakdown with the INDUCTOR specs listed for the actual part used on the eval board. Sure, there is tons of info about PICKING an inductor, but we shouldn't need to PICK anything, since there is an inductor already on the board, just tell us WHAT THE HELL IT'S RATINGS ARE, so we can know what current level the board can run with it's built in inductor.

Without this info, it's impossible to know if it's safe to reconfigure the board for 800mA. because if you drive the inductor into core saturation it can melt right off the board (after the high current destroys the ZD850 and your test LED's).

So again, thanks for the effort, Mark. This board does indeed look like it could be a real nice option for CPF type projects, because unlike some Eval Boards, it's not overpriced, or as big as a house, but a board-level hook-up diagram and spec info for the actual inductor on the board is still needed.

 

[macforsale]

*

 

[Luminescent]

Yes the formating is confusing.

Fig 4 is the 3 LED at 300-350 ma configuration and Fig 5 is the 3 LED at 1.5 amp. Values of R1 and L1 adjusted for current levels.

Fig 7 and 8 on page 11 give the connection information.

You are right about the inductor part number. It would be helpful to understand what the limits are. The 1.5 amp limit of the IC is the likely limiting factor. Review of the Bourns site for the inductor series might be helpful.

imageicon,
To get back to one of your earlier question. I would say the Eval Kit PCB is usable if you stay with 3 LEDs @ 350ma and keep within the voltage input limits.

Two things come to mind. One is that some drivers don't like open circuits and I don't know if this is one of them. The other is the external connections will be to surface mount pads and not thru-hole pads. Thru-hole would be better for wires, you will just have to be careful with pads. Easy to pull the pad off the PCB while still hot from soldering. The heat briefly weakens the copper foil to PCB base material bond. After it cools it's OK. It's still easy with the wire to pull the pad of the PCB, just be careful.

mike

Hi mike, I think you missed my earlyer point. Download the PDF doc again, all the figure numbers DID change.

The originally posted link was to - AN020_ZD850EVB_MANUAL_07 30 07.pdf

(This matches your figure numbers)

The currently posted link points to - AN020_ZD850EVB_MANUAL_08 24 07.pdf

(notice the changed date)

Maybe you need to clear your browser cache and update the page to see the new link?

The 3 LED 350 mA. example that WAS figure 4 is NOW figure 5, and figure 7 and figure 8 on page 11 are NOW figure 9 and figure 10.

None of the above changes anything substantive about your comments, but it could be a little confusing if someone new to the thread doesn't know that there are two versions.

The figures you referenced on page 11 are just poorly drawn board views, so one would have just about as much luck figuring it out for oneself by just checking out the board with a 10x loop (since it only appears to be a simple two layer board and everything is right there up front).

Of course the chip level diagrams are already available in the ZD850 chip documentation, and if I have to guess at component values and trace circuit pads, then what EXACTLY is the 'EVB MANUAL' doing for me?

Answer NOTHING, it's pretty much USELESS.

Dispite the bad docs for the eval board, I agree that it would be wrong to let the documentation issues spook you and prevent you from considering this chip, which could be a very good performer once you get the connections sorted out.

I also agree that it should not be too hard to sort out these connections on your own, with very little danger that you will goof something up and blow the driver.

This is not a boost converter so accidentally breaking a connection to the LED looks like it will also open circuit the inductor removing drive (I would still be careful to have everything hooked up correctly with good solid solder connections first because some outputs don't like to be open circuited, so it just makes sense to be cautious)

As far as looking at the Bourns site goes, that's not very practical, because they list several Bourns inductor series, but we have NO ASSURANCE THAT THEY ACTUALLY USED ANY OF THEM.

Everyone lists big name suppliers like Bourns, TDK etc for inductors in their datasheet recommendations, but inductors tend to be pricey, so if you can't work the kind of huge volume deal that impresses these companies, you usually end up subbing a cheap generic inductor from some no name vendor, and this could have quite easily happened on the eval board itself.

I am sure the factory can supply this info though.

Again, I don't want to sour anybody on the ZD850 eval board, because at only about 10 bucks it could be a great option, and bad documentation is a pretty common problem that all engineers have to deal with every day (no one will remember or care once you get things figured out and it's working).

 

[imageicon]

hey guys ty for all the information! geesh no wonder a noob light maker like me cant figure it out

my application is bike lights
i have produced some lights 2x cree XR-E emitters some taskled bFlex boards driven off 9.6v nimh batteries!

and some of the guys i bike with really likes them one of them work at a machine shop so he can make some aluminium housings for em, and we would try to cut down the cost on making them by using cheaper components than what i used for my prototypes. so really some differend design ideas are okay.. it was just when i stumbled across this chip it seemed all i would need from the product text but indeed it wasnt so easy when i started reading the manual

cheers jesper

 

[Luminescent]

hey guys ty for all the information! geesh no wonder a noob light maker like me cant figure it out

my application is bike lights
i have produced some lights 2x cree XR-E emitters some taskled bFlex boards driven off 9.6v nimh batteries!

and some of the guys i bike with really likes them one of them work at a machine shop so he can make some aluminium housings for em, and we would try to cut down the cost on making them by using cheaper components than what i used for my prototypes. so really some differend design ideas are okay.. it was just when i stumbled across this chip it seemed all i would need from the product text but indeed it wasnt so easy when i started reading the manual

cheers jesper

What do you think about a high performance linear constant current driver?

At high current your LED's will drop up to about 3.4 volts each, so two times 3.4volts = 6.8volts.

To calculate efficiency we devide this by the supply voltage 6.8volts / 9.6volts = 70.8% efficiency for the (fully charged pack).

In a linear regulator the efficiency actually goes up as the batteries discharge and their voltage decreases, so for a battery pack which has been discharged to 1volt/cell, we hit a much higher efficiency:

8volts/9.6volts = 83% efficiency

With the shape of the NiMH discharge curve, the overall average efficiency will probably be close to 74% which is just about as good as some switchers.

If you would like to try a linear driver, I have a fairly simple circuit with only about a dozen parts, which can drive any current level you want to well beyond 2 amps, but for CREE or Rebel, or SSC P4 LED's I would keep the drive down to about 700mA (and to even handle this current level you will need fairly good heatsinking).

The circuit is designed to be extremely forgiving about substituting other low cost alternate parts, so you should be able to get all the components required for less than $5.

It would require some soldering skills to build the circuit on a small piece of electronic breadboard, but if you know someone handy in such matters, it isn't a particularly complicated circuit.

Everything should fit on a tiny 1x1.5 inch PCB (and SHOULD BE, in order to keep the connections reasonably short). A small heatsink should be used on the MosFet Q1. If the same heatsink is used for Q1 as for the LED emitters, then the TO-220 metal tab on MosFet should be electrically isolated with thermal washer (anyone who has worked with this type of device will understand how to properly isolate the device from the heatsink).

Here are some additional notes:
All resistor 1/4 watt
Capacitors are low voltage monolithic, rated 25 volts or higher
Q1 is a T0-220 power mosfet. The IRF540 can be substituted with any 24Volt to 100Volt low RDSon N-Channel MosFet type.
U1 is not indicated but can be any low power opamp suitable for operation from 5 to 6 volts DC which is stable at a gain of 1 [internally compensated, unity gain stable] and that has a "Common Mode Input Range" which includes the negative supply rail [ground in this case]. Characteristics similar to MC33171 should be fine.

Here is the circuit -

Note: This example circuit is set up for a current drive level of about 730mA, which close to the maximum at which you can safely drive Rebel and CREE LED's, to drop this down to 350ma. cut the value of R3 and R5 in half to 5k.

 

[imageicon]

would prefer an off the shelf product thou
i will go support goerge at taskled i fancy his drives... and the guys who im making lights for will just have to pay thoose extra 20$, that is if you dont know of any linear controllers off the shelf thats cheaper...

i really admire you elektronic guys my lvl is like i know how to solder it and i know how to follow a plan but i cant do the calculations or the setups my self

 

[Luminescent]

would prefer an off the shelf product thou
i will go support goerge at taskled i fancy his drives...

I thought that this might be the case, but I thought I would show you what the circuit would look like, so you could see what the level of complexity would be. With only about a dozen components, your average ham radio operator or other electronic hobbyist could tinker something like this together in about ten minutes, but if you have never tackled something that looks like this before, it would definitely be better to stay with a commercial product.

Sorry but it doesn't look like the ZD850 would be a very good choice at 9.6 volts either. As your batteries discharge they drop down to about 1 volt/cell (or about 8 volts in this case), and with 2 LEDs in series the ZD850 seems to need 10VDC or more to drive it, so maybe one of the TaskLED boards would indeed be your best bet despite the higher price if you already know that one of their BUCK drivers would be ok with TWO LEDs at this 8 to 9.6 volt drive voltage.

 

[imageicon]

well actually i have handled diagrams and making of boards some years ago but i dont really have components or tools and when you say local i think the nearest elektronik store is 50km away...
and for my application wich will recive a pretty rough treatment on the mountainbike i like drives that is premade seems more durable

but ty for your help... i threw a volt meter on my batts and they start out at 10.6 or something and your right after an hours or 2 use they are noticely dimmer than when i plug it on but still plenty of light for my use.

i think ill change thou to 3x cree and then serie up 2 batts to give it 19,2v should ensure max brightness for the full duration.:devil:

 

[bikerjay]

I am now looking at this driver as well. Listed as $8.08 on the cutter.com.au site now. Thats hard to pass by. I am planning on driving 2 cree xp-e and want to use a 2cell Lion battery pack. The xp-e is rated at 3.5vf @700ma. The 7.4V nominal battery and forward voltage of 7V seems really awkward. Best solution I can come up with is to wire them in parallel, and hope they are well matched. Maybe I will just bump the drive current down a little to 1300ma. So I would be using this part with An input voltage of 6-8.4v and a current of 1300ma and a vf of ~3.5. I am guessing the zd850mr comes with the wrong inductor and the wrong resistor values, and I will have to change them. This could be difficult for me since I have never done surface mount before. Additionally I would have to calculate values based on that cryptic data sheet and source the parts which could be expensive. Can some one suggest some part values, or a different regulator all together. The data sheet seems to suggest a 22uH inductor. I am not sure about what resistor value to use for 1 led and 1300-1400mah current. I am set on 2 cree xp-e and a 2cell lion battery. Can a SOB do 1300ma off of 7.2V ?