S-BEC problems

[Giz]

I have just finished an aeroplane which uses an S-BEC and it is giving me lots (huge amounts!) of glitching problems. Has anyone else had an experience like this?

I did a radio check yesterday, aerial down and model on the ground, AND I could only get 3 metres away before all channels started fluttering. With one element of the aerial raised I could get about 8 metres away before the glitching started. With the aerial fully extended it appeared to be OK but a flight test (stupid, I know :eek: ) soon showed that it wasn't.

Replacing the S-BEC witha receiver pack caused the problem to go away and my aerial down range check was good at 60 metres+. No doubt about the culprit then.

I have emailed Aircraft Dave to see what he says but I am interested in anyone elses experiences because I don't want to replace it if the new one is going to cause problems too.

[GordonTarling]

Where in the model do you have the S-BEC located? If it's anywhere near the receiver or aerial, then this could be the cause of your problem.

[Giz]

Gordon,

The S-BEC is about 10 inches forward of the receiver. I had to add a 6 inch extension cable between the S-BEC and Rx to make it reach.

The basic layout is as follows:

Front of Plane - Motor - 3 inches - ESC - 3 inches - flight pack and S-BEC - 3 inches - Rx with aerial going to tail.

The S-BEC is alongside the pack.

[Giz]

Gordon,

One more thing. I just tried another test. With the Rx pack connected and the S-BEC still connected to the pack and switched ON, I get no glitching. I only get glitching when the S-BEC is directly connected to the Rx. So my conclusion is that the interference is not Rf generated from the S-BEC but is noise on the S-BEC to Rx connection. If this is the case, it won't matter where the S-BEC is positioned.

What do you think?

[GregoryJ]

Giz, you have a faulty unit. Send it back for a replacement. It sounds like there is noise on the 5V supply which should definitely not be there. There should only be about 10mV ripple which should give you zero interference.

Greg

[Steve_Bennington]

Heard a similar story a while back about UBECs..... seems these things use switching regulators (as opposed to an ESC's Linear regulator) which can kick out all sorts of RF interference....

Some people have used them without problems though.....

Regards,

Steve

[potifar]

one fix seems to be to wind about 3 turns of the x-bec to receiver cable on a ferrite toroid of proper characteristics.

the interference caused by the x-bec is conducted rf interference and is thus filtered.

if you guys don't have a source for those toroids i can talk to rc-direct and see if they want to make them available to distribution.

note that some high-frequency switching esc's cause a similar effect and are also aided by the addition of this toroid core filter.

peter berg

[JoeR]

Peter,

Do you have experience with the ferrite rings that Castle Creations provides? They are very small and get only one wrap of the signal wire on the ESC just before the reciever. I am just curious to hear what a receiver guy says about these. Thanks, Joe

[Steve_Bennington]

Peter,

Would this sort of interference be more likely to affect a Dual Conversion Receiver than a single conversion one? My Dualcon Rx's with brushless ESCS will only give me a good range check if the RX is foil wrapped!
We fly on 35Mhz here.

Regards,

Steve

[GregoryJ]

x-bec's are designed to operate in conditions where linear regulators would fail. For example, SBEC can drive 6 servos with an input voltage up to 40 NiCd cells. There is no linear regulator in the world that even comes close. Modern switch mode regulators have surprisingly low EMI. When problems like the one listed here arise, people automatically think its switch-mode BECs in general that have this problem. Not true. You should get a range check comparable to using an Rx pack regardless frequency (unless you're on a non-standard one), otherwise the unit is probably faulty.

[Andy W]

As explained, these BEC's have been painfully designed to ensure they do NOT emit RF interference that can impact the radio control of the model. Every now and then, however, a faulty unit can show up. Better to get it replaced, than fool around trying to figure out how to clean up the noise..
..a

[vintage1]

One trick apart from wrapping the leads in a torioid is to put a suppressor cap across the receiver +ve and -ve using a spare servo output channel to pick them up.

BUT the S-bec should already have suppression. Probably worth trying a new one first.

[Steve_Bennington]

Yep,

Worth trying a new one first....
The trouble with this place is that both manufacturer and sponseree will also post only good things about this product.

I'm stuck in my old radio ham teachings that switched mode supplies although cheaper, are always inferior to Regulated supplies...

Perhaps whoever thinks that 40v cannot be linear regulated to 5v at current, knows nothing about our telephone network!!!!!!!

Regards,

Steve

[Steve_Bennington]

PS. Vintage1, are you going to Binham LMA do tomorrow?

[vintage1]

what binham LMA?

[Steve_Bennington]

On Today, At Binham Airfield, Norfolk

Regards,

Steve

[John E]

I have been testing S-BEC in Bonnie for 3 months now with absolutely no glitching. The cable from the S-BEC to the receiver travels around the radio bay, no problem. I've never had a complaint from a customer either - we now have about 20 in use in UK.

It is interesting that Kontronik say that the ferrite ring fitted to their Jazz ESCs must not be removed.

Giz, I think that if the manufacturer says you have a faulty unit and offers to replace it, it's a good idea to take up the offer, and I'm sure you have.

Best regards,

John

[Giz]

Yes, Greg has very quickly offered to send me a new unit. It should have been dispatched yesterday so I am hoping it will arrive this week. I don't know how long shipping from South Africa will take. Sounds like great service so far.

When I get the new S-BEC I will obviously let you know how I get on. It seems like I have just been unlucky this time. I have range checked and flown the aircraft with a 700mah, 3oz receiver pack so I have a good benchmark for comparison. On three flights today, I experienced 2 glitches but that seems to be about par for the course at our field.

Thanks for the feedback and thanks to Greg for offering to fix the problem for me.

[GregoryJ]

Giz, its part of the service.

Steve, getting 5V from 40V using a linear regulator supplying 2A will consume 70 Watts ((Vi-Vo)*I). Using SBEC to do the same thing will consume 1.7 Watts (Po-Po/n), where n=efficiency=85%). Its very difficult to dissipate 70 W unless you have a very large surface area or the regulator is dipped in liquid nitrogen - so in theory you're right, it CAN be done.

EDIT: Equation should be Po/n-Po.

[Steve_Bennington]

Gregory,

Rather than use liquid nitrogen, I settled on a 2 stage regulation process,

I use an LM7824, then drive 2* L4940V5s in parallel. 3 little TO 220 packages and the job is done, for under £8 (about $15)

Gives 3A output all day, without getting hot


Regards,

Steve

[GregoryJ]

Hi Steve, the equations Pi=Vi*Io, Po=Vo*Io and Pdiss (power dissipated) = Pi-Po still hold for your method. i.e. you'll still be dissipating 70W if you are drawing 2A (obviously you're drawing a lot less). You're simply increasing the surface area by using 2 regulators (better heat dissipation) and also reducing the current through each regulator so that the Pdiss per regulator is less (but the Pdiss total is STILL 70W), thereby extending the time before the device enters thermal shutdown. You sound like you know what going on, so you'll know that these devices enter thermal shutdown when the silicon substrate reaches 150ºC. You'll also know that the temperature coefficient for TO-220 packages (the regulator you're using) is 50ºC per watt - for every watt it dissipates it increases in temperature by 50ºC. Add a large heat sink and you get maybe 10ºC per watt (I'm being conservative - its probably closer to 15 or 20ºC/W). This of course means that each regulator (you're using 2, right?) can only dissipate 15W before reaching thermal shutdown temperature. Translated to current, that's only 0.375A each or 700mA in total before it shuts down. Not to mention the power you're wasting that could have been used for the motor, servos, receiver, gyro, etc.

In contrast, a switched regulator, or SBEC at least, dissipates just 600 mW (0.6W) and increases in temperature by 20ºC under the same conditions where your linear regulator dissipates 30W and increases by 150ºC - I'm talking about the substrate of course. You'll realise that I'm assuming ambient temp is 0ºC. In more realistic temps (Norfolk is about 10 to 15ºC this time of year?), 28W will cause it to reach shutdown temperature.

Some of the guys over on the electric heli forum have done actual measurements on a 4 servo heli and it peaks at 2.5A which is on average about 600mA per servo (quite acceptable I think). You can easily work out that a linear regulator like yours driving 3 servos (aileron, elevator, rudder) will enter thermal shutdown if the input (battery) voltage is above 13.3V. So good luck if you're using more than 12 cells.

I really don't mean to start a battle between linear regulators or switched regulators (been there, done that, got the t-shirt) but the equations do speak for themselves.

Regards
Greg

EDIT: Correction, I see its 3 TO-220's you're using. So at 10ºC/W, its 15W per regulator, with a total of 45W before thermal shutdown = 1.125A. So the bit that says "In contrast..." should read (changes in bold):

In contrast, a switched regulator, or SBEC at least, dissipates just 1W and increases in temperature by 33ºC under the same conditions where your linear regulator dissipates 45W and increases by 150ºC - I'm talking about the substrate of course. You'll realise that I'm assuming ambient temp is 0ºC. In more realistic temps (Norfolk is about 10 to 15ºC this time of year?), 42W will cause it to reach shutdown temperature.

[Steve_Bennington]

Fascinating....

Reminds me of a white paper I read once proving that bumblebees can't fly...

My quick solution works, and doesn't get hot, and is also very cheap, and whats more, does not generate ANY discernable RFI or ripple...
I'm also drawing 3A constant at 5v, for an hour... still no shutdown....

Thanks for the detailed theoretical explanation though!

Regards,

Steve

[headless]

Curious; what input voltage (and from what type of source) are you providing, Steve?

[vintage1]

I reckon hes running it off a separate 4.8v receiver pack :)

I spent my life designing linear power amps, and believe me, you cant escape those watts figures. If you think you are, something else is going on :)


Trouble with this SBEC lark is by the time you get up to a voltage large enough to warrant it, and think about teh weight of teh sbec, its pretty much that you can affod the weight of a 2sLIPO pack and a 7805 for receiver and servos anyway.

You can't do a switching regulator without a choke, and chokes are heavy things.

[pmackenzie]

Just before the UBEC came along a friend with a LOGO 10 wanted a BEC for 14 cells. Below is what I made for him, although in his case we put it on a heat sink.
There is one regulator to step down to 10 volts, then a final one to 6 volts. This splits the heat evenly between each regulator.
He wouldn't even consider a rcvr pack since all of his prior experience was with electrics with a BEC. He is still using the linear regulator with no problems.
The big advantage of a BEC over a rcvr battery is that it is always "charged",will never go dead, and is lighter than even a 250mahr 4 cell pack. A two cell LIPO pack will still need a regulator, plus a charger.
I have both a UBEC (14 cells/4S lipo) and an SBEC (22 cells/7S lipo)
in heli's. No problems with either.
The UBEC with wires weighs 23 grams. The SBEC is tiny and only weighs 8.5 grams!
Pat MacKenzie
P.S. I use the one shown in a 10 cell ECO, just so I could use the OPTO controller I had.

[GregoryJ]

I reckon hes running it off a separate 4.8v receiver pack
Yeah, refusing to accept or believe the power equations doesn't make them dissapear.

You can't do a switching regulator without a choke, and chokes are heavy things.
You can use smaller components with higher switching frequencies to achieve the same power as lower frequencies with larger components.

There are merits for using linear supplies, but there are power dissipation issues that need to be considered which are affected by input voltage and supply current. Vintage1, on a 2 cell LiPo pack (7.2V), I agree, a linear reg is best. Take it to 3 cells (10.8V) and see what happens. A 7805 on a decent heat sink (30ºC/W) will only be able to dissipate about 4.5W before thermal shutdown (ambient = 15ºC). At 10.8V input, the max current is only 780mA for rx, ESC electronics and servos. It may work on a 2 servo model but its pretty risky.

Greg

[leccyflyer]

Since this thread has developed into a discussion of electronics and the original query appears to have been resolved I'm going to move it to Modelling Science.

Brian

[vintage1]

Wel I have had one regulator shut down on me on 3s LIPO on a 4 servo model.
ACTUALLY maybe I have had two, but that was a loing time ago and I was a real newbie then. Only part newbie now :)

Staistically it depends on cooling, heatsinks, and servo activity.

Its like when I was desgning power amps. In the real world you could get away with power levels that you didn't dare go for on the bench.

Until some stupid DJ ran them up flat out for 4 hours in a hot S. African club :)

(seeing that you are frm the Cape, greg, I used to work for a company called Audio Geometry in Jo'burg, for a while :))

[GregoryJ]

Staistically it depends on cooling, heatsinks, and servo activity.

True. And it consumes precious battery enery. I was thinking about it. In the example with a 40V supply supplying 2A, the unit wastes 70W. Most medium sized heli's hover on 100W per kg. So the power wasted is eqivalent to the power used to hover a 700g heli! Even if one had sufficient cooling, linear regs are exceptionally wasteful.

Until some stupid DJ ran them up flat out for 4 hours in a hot S. African club
He must have been popular :-)

Did you ever make to Cape Town? I spent a few years living in London and I'm REALLY glad to back in CT. Recognise this?

[splash99]

Greg , just ordered 4 SBECS from Firmtronics.
The SBEC has been pointed out to me yesterday in one of the Forums, and I have been looking at suppliers.
Your explanations on how these gizmos work (not that I really understand it;) ) and how you dealt with the original problem of Giz in this thread just got you a new customer.

[GregoryJ]

Thanks for your support Splash, new customers are always appreciated. I'm glad you find this info useful. I think there are many more who are unaware of the issues around linear regulators.

With relatively new products like this, there's always skepticism, especially with products that play such a crucial role in keeping your expensive investment (or liability, whichever way you look at it) in control. I was pleased to see SBEC being put through its paces on some larger electric helis in the USA (probably the most demanding application). These heli's are peaking over 3500 W (over 4 horse power) for drag racing (which are now beating their gas counterparts!!). If you're interested, there's more info here: http://www.rcgroups.com/forums/showthread.php?p=2155489#post2155489

Regards,
Greg

[vintage1]

Did you ever make to Cape Town? I spent a few years living in London and I'm REALLY glad to back in CT. Recognise this?

Definitely. My sister was living in stellenbosch at the time, and we did the garden route down from Durbs, camping all the way, and then 'hailed @ss' back through teh Karoo and the Vrai Staat to get back to Joeys in one day....

I've been back since to Kwa Zulu. Lovely experience till we got back to Durban on the return leg...and even worse re-visiting a friend in Jo'burg...8ft high razor wire everywhere ....

[Giz]

Okay, here is the conclusion to the S-BEC problems story. I have been holding off from posting this for a while until there was a final resolution.

Greg at Firmtronics in South Africa sent me a replacement S-BEC very quickly but it suffered from exactly the same problem as the first one :eek: . Greg was very surprised and decided to test another one out personally before sending it to me. During this try-out testing, he found the same problem as me with some receivers. So. it turns out that the problem occurs with just some receivers and not others and was not a faulty S-BEC as such. Greg identified the problem as conducted RF interference from the S-BEC that was not being filtered by the receiver.

Greg looked into the problem and tried a number of options to improve the output filtering on the S-BEC but they showed little improvement. Having re-read all of your posts, I decided to try a ferrite ring on the power lead. The only one I had was from my Schulze 6-330 charger. I found that if I used a servo extension lead and wound the cable through the core six times that my range test improved from 3-5m to 25-35m. Quite dramatic!

I considered 25-35m to be a reasonable aerial-down range test and decided to fly. I found no glitching problems at all and have been flying the aeroplane regulary for the last two months.

Greg has been working on the ferrite rings too, trying to identify the most efficient configuration and has now succeeded in finding an effective, lightweight solution.

So I think the issue is finally closed. In conclusion, some receivers are very sensitive to the ripple or interference from the S-BEC whilst others do not show the problem at all. The problem can be completely eliminated by using a suitable ferrite ring on the power cable, wound a suitable number of times.

Below is a picture of my ferrite ring wound six times. Thanks to everyone for their advice and to Greg his quick response and for keeping me in the loop during the investigations.

[Larry3215]

Thanks for the update Giz.

Which recievers are affected and which seem to do well?

Larry

[JoeR]

Can you list the receivers/brands that are problematic with the SBEC? I have one of these to put in a Logo10 helicopter and don't want to risk glitches. I have a ferrite ring that came with my Shulze charger like the one mentioned above and it is a monster! Anything more appropriately sized for the application?

[Giz]

I can't give you a list as I have only tried the S-BEC with one receiver which was a Hitec Electron 6. This is the one that had the problem.

[KatManDEW]

I've been traveling down the same road myself. I just got some ferrite rings last night. I tried three different brand new receivers and couldn't get anything close to a good range check. I tried a FMA FS5, FMA M5, and Electron 6 receiver. And I tried two different SBEC's, a 5 volt and a 6 volt. The strange thing is that I have a SBEC in another airplane with a FMA FS5 receiver and it range checks fine :confused:

[splash99]

I see you wind the power cables and the RX-leads around the same ferrite ring - I never thought of that, asuming there would/could be some effect between the two but I have to confess that I don' t understand this subject....
Woudl be glad to hear from anyone who has knowledge if there might be interferrence or if the ferrite ring just smoothens out every - and anything?

[Larry3215]

It could be your esc causeing the trouble.

Ive had better luck putting ferrites on esc leads than bec leads. I use sbec and ubec and all Berg recievers - 5's and 6's and all Castle ESC's (10's 25's 35's 80's).

Most setups work fine. But it seems like every third plane has issues and needs ferrites. I reciently made up an extention with chokes to test, but O dont have a final opinion yet. I also purchaced an opto isolator from these guys

http://www.medusaproducts.com/Other/OPTO-001.HTM

I'll be testing that this weekend along with thier 2 amp BEC when it gets here.

The ferrites (I'm useing the small ones from Aircraft world) have been fairly effective for me in increasing range.

Larry

[headless]

interesting.

[Larry3215]

I just realised this got moved to modeling science.

I don't know if I agree with the move.

This is the kind of info I would have wanted to see in power systems since it involves bec's and esc and how to make them work well.

Larry

[dave morris]

I was just going to start a new thread when I happended upon this one. I am having the same problems, 3-5 yards and servos go nuts. Should I try for a replacement? Not the ESC (I don't even bother to install it, just plug the sbec into the rx). Range checks out fine with Rx pack (triple the distance easy). I will try the torrids if that is the recommended fix (assume the sbec is not faulty). Just to get it right, I should wrap the sbec output thru and around the torrid? The input lead I assume is fine. The setup is a 5s lipo and both a hitec 555 and a berg5ch, both have very poor range test (fine with rx pack). The sbec is located about 5 inches away from the Rx, the Rx ant lead goes no closer. The sbec is located near a servo lead, perhaps that servo lead should get some chokes or a torrid? I am really stumped, not to mention grounded with my new toy.

[Giz]

Splash99 - only the wire that supplies the Rx with power from the S-BEC is wound around the ferrite ring. I removed the plug to do it and then refitted it. To get more turns, I added a 12 inch servo extension cable which is why there are two different types of wire showing.

Larry - it definately is not the ESC (Jeti 70 Opto). Supplying the power from a receiver pack but leaving the S-BEC in place and switched ON shows no problems. The ferrite that I'm using is 28mm long with 15mm outside diameter - a bit large but at least I can get six turns on it. I found less turns gave shorter range check.

Dave - I didn't find the position of the S-BEC was critical because the interference turned out to be conducted and not radiated. My range check with an Rx pack was 60-70m so if you are only getting 3 times 3-5m you may have other issues.

I think this thread should have been left in Power Systems which is why I put it there in the first place.

[splash99]

thank you Giz for clearing up the confusion - that makes sense to me again

[Larry3215]

Splash99 - only the wire that supplies the Rx with power from the S-BEC is wound around the ferrite ring. I removed the plug to do it and then refitted it. To get more turns, I added a 12 inch servo extension cable which is why there are two different types of wire showing.

Larry - it definately is not the ESC (Jeti 70 Opto). Supplying the power from a receiver pack but leaving the S-BEC in place and switched ON shows no problems. The ferrite that I'm using is 28mm long with 15mm outside diameter - a bit large but at least I can get six turns on it. I found less turns gave shorter range check.

Dave - I didn't find the position of the S-BEC was critical because the interference turned out to be conducted and not radiated. My range check with an Rx pack was 60-70m so if you are only getting 3 times 3-5m you may have other issues.

I think this thread should have been left in Power Systems which is why I put it there in the first place.

I was replying to KatManDEW. I wasn't sure he had done all the testing you did to eliminate the esc as the source.

I'm going to test the chokes and opto unit later this week and I'll let you know how good a job they do.

Larry

[vintage1]

Did putting a smoothing capacitor across a spare servo output (+ to -) on the RX make any difference?

That ought to make a huge difference WITH a choke, and may indeed be enough without it.

[Giz]

Vintage,

I did try a quick hack of sticking a capacitor supplied with a 400 motor across the power supply lead at the Rx end but I didn't get any improvement. Then again, I didn't know what I was doing and may have used a completely inappropriate capacitor.

I believe Greg tried this at the S-BEC end without much improvement. I don't know whether he tried this in combination with the choke. What specification of capacitor would you recommend?

[vintage1]

As you said, a suppression cap. 10-47nF. Maybe with an electrolytic as well.


If it didn't work it suggests a very low impedance ripple indeed.

Its possible to make up a gizmo comprising two RF chokes (1.5uh) and a capacitor to fit instead of that rather bulky toroid.

[Steve_Bennington]

Giz,

Now I'm confused!!!

You posted that same pic in Castle creations forum about a glitching ESC on 35mhz. I'm not sure if you were trying to say that you had not experienced glitching on 35Mhz, or that you had, but it was not the ESC, and the SBEC caused it.....

Did you Isolate the source of the glitching? CC ESC or SBEC??

This is the post in the CC forum...

http://www.rcgroups.com/forums/showpost.php?p=2342712&postcount=27


Regards,

Steve

[Giz]

Steve,

What I was saying was that the symptoms sounded very like the problem that I was having. I described my situation as a 'BEC' connected to a Hitec Electron 6 Rx.

I wasn't trying to be clever, just helpful and I the time the S-BEC investigation was still in progress. To clarify, the glitching was on 35Mhz and was caused by the combination of an S-BEC and Electron 6. The ESC in this case was a Jeti 70 Opto and therefore had no BEC.

There were two key points:

1) The receiver can be a significant contributing factor to the problem as some receivers seem to be suceptable whilst others are not.
2) You might be able to fix the problem with a ferrite core if the source is conducted interefernece.

[Steve_Bennington]

Giz,

That clears it up! Thanks!

Steve

[Larry3215]

I started a thread in Power system with some first results of range checking with ferrites, chokes and Opto coupler on SBEC and UBEC.

http://www.rcgroups.com/forums/showthread.php?p=2421955#post2421955

Larry