ULtra Compact 1800W charger + Eltek programming

Guys i think i have found the world most compact and powerfull charger ever 8)

It's a ELTEK 1.8kW flatpack S !!

These are availlable in 24 and 48V and take from 100 to 240V.

But what is fantastic is that they have the unbeatable 47 W/cu inch density !! and are 96.5% efficient!!!

These are the size of these cheap 2A 48V charger!!!

I will fit one or two in my NYX frame.

By modifying one i was able to shorten it to about 7.5" long!!

These are 2.8 x 1.6 x 8.5" size.

I contacted the Canadian representative of these and i will get a quote for more.. These are perfect to combine with the Adaptto.

They are OEM 53.5V and 33.7A and output 1800W on 240V ac input and about 1000W on 120V ac

I had no BANANA gus .. very sorry for that.. I only had apple

But see how small it is !!

I also atached a comparaison between some popular rectifiers ( server power supply we use on E-S)

Doc

wow that is incredible small.

And perfect for my 14s battery if it can charge up to 57V!!!

How much is it? Is it protected against vibration?

cwah said:

wow that is incredible small.

And perfect for my 14s battery if it can charge up to 57V!!!

How much is it? Is it protected against vibration?

Click to expand...

Yes it is small.. even it is difficult to believe that it output 1.8kw from being so small !

Yes it can do up to 57v but it need to be programmed.. Eltek said it is possible to get them preprogrammed at the desired voltage and current.

They are well built for damping vibration with silicone on the important ckmponents inside.

I am waiting for their response for price quote.

Doc

So at 30Amps of a 48V battery 54.6V charging it would only generate 60W of heat

Do you think the noise would be enough low to night charging on the same room?

It has variable rpm fan?

looks maybe the first real competitor to the satiator in terms of efficiency

where can we read the comparison between other popular rectifiers?

thanks for sharing!

Nobuo said:

So at 30Amps of a 48V battery 54.6V charging it would only generate 60W of heat

Do you think the noise would be enough low to night charging on the same room? NO

It has variable rpm fan? YES

looks maybe the first real competitor to the satiator in terms of efficiency

where can we read the comparison between other popular rectifiers? HERE by searching :wink:

thanks for sharing!

Click to expand...

Doc

Hi Doc !

Do you have any good photos of the connections ?

REgards
/Per

Very tiny... got the price cotation back Doc?

About 400$ each brand new :lol:

Doc

Doctorbass said:

About 400$ each brand new :lol:

Doc

Click to expand...

I was expecting a high price, but not that high. =/

hi Doc you Very helpful to all es . but one thing have you seen the size the Eltek Smartpack controller, it Big. if you want change amp and volts.
http://www.eltek.com/detail_products.epl?k1=25507&close=1&id=1123846

i bet it not cheap. (

i was try to find flat pack1 i only can flatpack 2 for $80

Should be just a matter of finding out the correct CAN message to send and have a little arduino to do that.
.....

Everytime I see something like this, I get all excited.. until I see it doesn't go up to 66v.

looks like I'm still on the doubled up 33v meanwells for now... once you get used to charging at 1300w, you dont want to go down.

eTrike said:

One of these just went for $160 on ebay

Click to expand...

Sorry, that was me

Just got it delivered today

I plan on using it to charge my 2014 Zero SR, but still need a second one to get the voltage high enough.
I already have a fast charger for the Zero consisting of 2x TDK FPS1000 PSU's but looking for more power in a smaller package
The FPS1000 is exactly the dimensions and the weight of the meanwell RSP-2000, So a transition to this Flatpack S would double the power and more than half the weight and even more decrease the volume for me.

Connecting the PSU to get it to output is super easy. I'll take a picture of the correct connections tomorrow. Just connecting the 230V volt supply was enough to get it to output 53.5 Volt. No further connections necessary.

If you look at the back side connections there are 5 ports. the 2 ports on the left each have 6 spring like connections (tabs) and all 6 tabs are connected together. These 2 leftmost ports are for the AC supply input (L1 and L2)
the 2 rightmost ports are also 6 tabs each with all 6 connected together per port and these are the DC output connections (minus on second to right port and plus on the rightmost port)
The middle port is a little special as it has 3 tabs on the bottom (All 3 are ground for the AC supply) and the top 4 are yet unknown to me. Probably the CAN/RS-485 and possibly 5 of 12V. I'll try to find out tomorrow as well.

53.5 Volt times 2 is not enough to get Zero to full charge so the output has to be altered via the RS485 bus. I read somewhere that that is a one time process. The PSU will keep it's output voltage after that. So the chase for the correct CAN-bus message to set it to 57.6 volt (maximum output voltage) is ON. :lol:

The picture below shows how the connections are made in a normal 19" rack (ATTENTION, the picture is from a flatpack 2 NOT a flatpack S but the connector is almost the same). The first thing i noticed is that the DC outputs are reversed compared to my flatpack S. Also in the picture it shows 3 connections on the top for the CAN-bus where the Flatpack S has 4 tabs.

Well, I had some success connecting to the Flatpack S

Of the middle port (with the 4 tabs on the top) the 2 tabs on the top right of that port are for the CAN-bus (not RS485!)
The top right pin (of the middle port) is the CAN-H and the pin to the left of it is the CAN-L.

That's consistent with the picture in my previous post if disregard the fourth pin in the port.

Connecting it all up to an arduino with a can-bus converter it took some time to discover that the bus speed is 125 kbps which is kind of an odd speed. :?

Anyway it outputs this message about every 2 seconds :
32 (can bus ID) 27 20 20 113 17 8 32 0
and every 15 seconds
0 (can bus ID) 20 20 113 17 8 32 5 0

The second message looks a lot like the first but shifted 1 byte.

When I put a load on the PSU (4 lightbulbs 12V 21W, measured at 1.90 A) the last number of the first line reads 12 instead of 0 and to me that looks like the current draw.
roughly calculating that 12 out of 255 (8 bits) equals a range of 0-40A
The second message however doesn't change at all.

So with 1.90 Amps the message every 2 seconds
32 (can bus ID) 27 20 20 113 17 8 32 12
and every 15 seconds
0 (can bus ID) 20 20 113 17 8 32 5 0

Next was trying to get a temperature reading but no change whatsoever from the values.

I could imagine that the 113 value is the voltage on a range of 0-120 Volt (113/255*120 = 53.2) but I haven't got a load large enough for the voltage to drop.

remmie1972 said:

Well, I had some success connecting to the Flatpack S

Of the middle port (with the 4 tabs on the top) the 2 tabs on the top right of that port are for the CAN-bus (not RS485!)
The top right pin (of the middle port) is the CAN-H and the pin to the left of it is the CAN-L.

That's consistent with the picture in my previous post if disregard the fourth pin in the port.

Connecting it all up to an arduino with a can-bus converter it took some time to discover that the bus speed is 125 kbps which is kind of an odd speed. :?

Anyway it outputs this message about every 2 seconds :
32 (can bus ID) 27 20 20 113 17 8 32 0
and every 15 seconds
0 (can bus ID) 20 20 113 17 8 32 5 0

The second message looks a lot like the first but shifted 1 byte.

When I put a load on the PSU (4 lightbulbs 12V 21W, measured at 1.90 A) the last number of the first line reads 12 instead of 0 and to me that looks like the current draw.
roughly calculating that 12 out of 255 (8 bits) equals a range of 0-40A
The second message however doesn't change at all.

So with 1.90 Amps the message every 2 seconds
32 (can bus ID) 27 20 20 113 17 8 32 12
and every 15 seconds
0 (can bus ID) 20 20 113 17 8 32 5 0

Next was trying to get a temperature reading but no change whatsoever from the values.

I could imagine that the 113 value is the voltage on a range of 0-120 Volt (113/255*120 = 53.2) but I haven't got a load large enough for the voltage to drop.

Click to expand...

Welcome to the forum remmie!

Nice tests!.. What would be awsome is to find a way to reprogram the set voltage to the desired value. These can be programmed for 58Vdc witch wold be awsome for the Zero motorcycle battery of 116V. That would only require 2 of these in serie to put 3600W in a very ultra compact size for charging the battery!

I guess it would be possible to modify the resistor divider that measure the voltage to false the reading and get the 116V, but let see if someone can get it reprogrammed!

I already asked to the Canadian sales rep from Eltek and he said he would help me to hget the software but he can't....

Doc

Doctorbass said:

Nice tests!.. What would be awsome is to find a way to reprogram the set voltage to the desired value. These can be programmed for 58Vdc witch wold be awsome for the Zero motorcycle battery of 116V. That would only require 2 of these in serie to put 3600W in a very ultra compact size for charging the battery!

Click to expand...

you would need 4 chargers to charge 3600w@116v. not two

izeman said:

Doctorbass said:

Nice tests!.. What would be awsome is to find a way to reprogram the set voltage to the desired value. These can be programmed for 58Vdc witch wold be awsome for the Zero motorcycle battery of 116V. That would only require 2 of these in serie to put 3600W in a very ultra compact size for charging the battery!

Click to expand...

you would need 4 chargers to charge 3600w@116v. not two

Click to expand...

No. Same amperage, twice the voltage..... doubles the wattage.

MrDude_1 said:

izeman said:

Doctorbass said:

Nice tests!.. What would be awsome is to find a way to reprogram the set voltage to the desired value. These can be programmed for 58Vdc witch wold be awsome for the Zero motorcycle battery of 116V. That would only require 2 of these in serie to put 3600W in a very ultra compact size for charging the battery!

Click to expand...

you would need 4 chargers to charge 3600w@116v. not two

Click to expand...

No. Same amperage, twice the voltage..... doubles the wattage.

Click to expand...

sorry. i don't get it. the charger does 1800w@58v. two of them can do 3600w@58v or 1800w@116v.
do you say you want to double the voltage and keep the amps still at ~30? how would that work? so it would turn into a 3600w charger?! i'm confused.

izeman said:

sorry. i don't get it. the charger does 1800w@58v. two of them can do 3600w@58v or 1800w@116v.
do you say you want to double the voltage and keep the amps still at ~30? how would that work? so it would turn into a 3600w charger?! i'm confused.

Click to expand...

wattage is not "at" any specific voltage. You take the Amps, and multiply them by the voltage (or visa versa) and you get wattage.
so placing them in series you keep the amperage the same, but you double the voltage... making twice the wattage.
placing them in parallel, you double the amperage, but the voltage remains the same... making twice the wattage.

doing what hes talking about,
1 of them is 1800 watts. at 58v dc that would be 1800/58 for 31.03 amps

putting two of them in series, its 58v +58v for 116V... but the same amperage... it remains at 31.03 amps.

116*31.03= ~3600

2 of

pm_dawn said:

Should be just a matter of finding out the correct CAN message to send and have a little arduino to do that.
.....

Click to expand...

There is no 'correct message'. It actually has a quite complex communications protocol, with logging in, pinging, selftests, etc. And protocol is not publicly available.
Your best bet would be to buy a complete system with controller, tap in to the line and reverse engineer it. Not for average people for sure, and not for most of us here. But doable.

circuit said:

pm_dawn said:

Should be just a matter of finding out the correct CAN message to send and have a little arduino to do that.
.....

Click to expand...

There is no 'correct message'. It actually has a quite complex communications protocol, with logging in, pinging, selftests, etc. And protocol is not publicly available.
Your best bet would be to buy a complete system with controller, tap in to the line and reverse engineer it. Not for average people for sure, and not for most of us here. But doable.

Click to expand...

Yeah, assuming minimal security a programmer makes cracking it much easier... almost fun.
You can trace the messages down by programming it at various voltages, then when you understand that part, you can just send your own message.

but... at that point you may still be unable to program it as high as you want, as it may be scaled so that you can only go so high... or it may be more secure (unlikely)... etc.

I would just find a friendly engineer that works for them, and talk to them directly. No sales guys.

MrDude_1 said:

I would just find a friendly engineer that works for them, and talk to them directly. No sales guys.

Click to expand...

Let me know when you do...

what a nice find DOC! Put it into a peli 1170 case together with the Adaptto coil and you have a very transport friendly charger.
do you know if eltek also offers lower power, but smaller size 48V psu (as ex 1000W)? The 1000W flatpack 2 charger has same size as 1800W model.

madin88 said:

what a nice find DOC! Put it into a peli 1170 case together with the Adaptto coil and you have a very transport friendly charger.
do you know if eltek also offers lower power, but smaller size 48V psu (as ex 1000W)? The 1000W flatpack 2 charger has same size as 1800W model.

Click to expand...

yes they have! they have 1000W same size i think.

Doc

circuit said:

MrDude_1 said:

I would just find a friendly engineer that works for them, and talk to them directly. No sales guys.

Click to expand...

Let me know when you do...

Click to expand...

i can just see it now. engineer spends 4 months writing code and finally has a successful product and some dufous wants him to spill the beans secretly so he can hack and sell his own design. yeh sure, i know a lotta engineers who would gladly do that. NOT.