Widebands That You Would or Wouldn't Buy

Discuss MegaSquirt, VEMS and other non-free hardware and software here.
Post Reply
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Widebands That You Would or Wouldn't Buy

Post by Fred »

In a similar style to the Standalones That You Would Never Use thread, this is going to be a feedback based wideband shootout!

Unlike the other thread, I'll maintain two lists, a positive list, and a negative list. Feedback for both welcome. Currently available items only, or those that were available when they went up in this thread.


Widebands you WOULD buy and use:
  • 3 - Innovate LC-1/LM-1/etc [Please read the full discussion, not all innovate products are equal, the best are not that great, and the worst are bad]
  • 2 - 14point7.com Spartan
  • 2 - 14point7.com SLC series
  • 1 - Innovate LC-2/etc
  • 1 - PLX
  • 1 - AEM UEGO
  • 1 - NGK - AFX
  • 1 - NGK - AFRM
  • Items above this line are safe bets!
  • 0 - Zeitronix
  • Items below this line best avoided!
  • 0 - Prosport
  • 0 - Tech Edge
  • 0 - Ecotrons ALM
  • 0 - betterpartsx BPSX

Widebands you would NOT buy and use:
  • 2 - betterpartsx BPSX
  • 2 - AEM UEGO
  • 1 - Ecotrons ALM
  • 1 - Tech Edge
  • 1 - NGK - AFX
  • 1 - Prosport
  • 1 - Innovate LC-2/etc
  • 1 - Innovate LC-1/LM-1/etc
  • Items above this line best avoided!
  • 0 - Zeitronix
  • Items below this line are safe bets!
  • 1 - NGK - AFRM
  • 0 - PLX
  • 0 - 14point7.com SLC series
  • 0 - 14point7.com Spartan
If results start to become clear, I may add an ordered list below. Items between the dividers received no votes in either list, or equal votes in both lists, and should be considered "unknown".

See the ECU thread for info on how to interpret these and possible caveats.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
toalan
Wideband Wizard
Posts: 241
Joined: Tue Jun 10, 2008 2:53 am
Location: Toronto Canada
Contact:

Re: Widebands That You Would or Wouldn't Buy

Post by toalan »

I am speaking not as the guy who designs and sells closed sourced widebands @ 14point7.com, but as the guy who made the open source wideband SLC Free, viewtopic.php?f=16&t=2083

This is my personal experience with other brands of lambda controllers

Innovate LC1- Tested 1 unit. It is accurate and fast, uses a mix of commercial and industrial grade components. Sensor life is very short and requires constant calibration. I believe the issue with the LC1 is that it does not properly detect the sensor temperature and that is the source of most of it's problems. Has over voltage and reverse polarity protection but no fuse. Uses their own uC + firmware for lambda control. Uses Bosch 4.2

PLX SM AFR- Tested 1 unit. It is accurate and fast, uses a mix of commercial and industrial grade components, creates alot of heat which is problematic since the DACs they use is commercial temperature grade. Uses low end electrolytics, which again is a big problem due to the high heat output of the unit. Has absolutely no protection on the power supply, no over voltage, no fuse, no reverse polarity protection. Uses their own uC + firmware for lambda control. Uses Bosch 4.2

BPSX- tested 3 units, It is not a wideband, I do not know what it is, but you are better off with a narrowband sensor + narrowband gauge. Uses their own uC + firmware for lambda control/miscontrol. Uses Bosch 4.2

NTK AFX + NTK sensor- tested 1 unit, very well built, has good protection circuitry, I believe they use all industrial grade components. When tested with 11.78 AFR test gas, the unit read bout 0.5 AFR richer, both the unit and sensor was brand new. Uses NTK ASICs for lambda control. Uses NTK lambda sensor

AEM (the one that is a gauge)- tested 1 unit (unit was used, sensor was brand new). I do not recall the internals but I also do not recall being impressed or dissapointed with it so the internals must be around average. Tested with 11.78 AFR gas, the unit read 0.5 AFR richer, the unit was used but the sensor was brand new. Uses Bosch ASICs for lambda control. Uses the Bosch 4.2

Glowshift/Prosport- test 2 units, both units were new with new sensor. Internals are good, uses industrial grade components, has proper voltage protection. When tested with 11.78 test gas both units read about 1 AFR richer. Uses Bosch ASICs for lambda control. Uses the Bosch 4.9

Tech edge 2Y1- Accurate, very slow. Internals are good, industrial temperature components, has proper voltage protection. Uses their own uC + firmware lambda control. I tested it with the Bosch 4.2, 4.9 is also supported but I have not tested.

Tech edge 2J2- Accurate, very slow. Internals are good, industrial temperature components, has proper voltage protection. Uses their own uC + firmware lambda control. I tested it with the Bosch 4.2, 4.9 is also supported but I have not tested.

Overall the most accurate units tend to use controllers developed from scratch, ie do not rely on Bosch/NTK Asics.

My personal criteria in order of importance is; accuracy, reliability, and then response time. If my 14point7 controllers are not available, I would reccommend Techedge units, after that it is a tie between Innovate and PLX. I personally dislike the owner of techedge, so I am not playing favorites by putting Techedge above the rest of the competition.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Widebands That You Would or Wouldn't Buy

Post by Fred »

I have a question, when you say "accurate" how accurate meats the definition of "accurate", when you say inaccurate you give a number eg 0.5 or 1.0 (:-o!!!), so is accurate <0.1 or <0.01 or?

May I interpret your post to place a "would buy" against your own, techedge, innovate and plx, a "would not" against the others you listed, and nothing against any others that are there (if any)?

Another question, when you say you value reliability, do you mean failure or do you mean consistency of results?

Personally I want it to be accurate and fast, and not fail, in that order. If it does fail, that's OK, as long as I know about it and can replace it with a superior device. Having it slow is a huge problem while trying to tune based on logs with any dynamic content. Have you ever driven anywhere without moving your right foot for significant periods of time? I doubt it :-)

With respect to accuracy, on the rich side, I'd want it to be wrong saying leaner than reality. Any amount richer than reality invites you to tune your car leaner than you think, and melt things. On that basis 0.5 AFR richer than reality is an utter utter fail for me.

FWIW, I only put current things in the list. If I was putting older stuff, your old JAW units would have got a negative vote. No favourites here either :-)

I really value and appreciate your input on this! Thanks :-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
JaniM
QFP80 - Contributor
Posts: 35
Joined: Fri Feb 15, 2013 6:13 am
Location: Finland

Re: Widebands That You Would or Wouldn't Buy

Post by JaniM »

One candidate Prosport, LSU4.9
http://www.prosport.com.tw/
Hell Fire
DIP8 - Involved
Posts: 24
Joined: Wed Jul 31, 2013 11:12 am

Re: Widebands That You Would or Wouldn't Buy

Post by Hell Fire »

I had the TechEdge 2Y2, if I had the choice to buy again, I would have chosen differently. Good enough controller, decent parts used, but it's response time is quite awful, the amount of things they try to cram on to a ATmega168 is quite bad, adds a lot of complexity to the circuit (think multiplexed IOs), and the support from TechEdge is terrible if you can get to them at all.

The firmware is all oldschool assembly programmed and not all that good (I discovered it was highly lacking in error/race condition handling) and doesn't lend itself to being updated by TE very often. It's an old product trying to do too much all in one box and it shows.
Last edited by Hell Fire on Fri Sep 06, 2013 11:59 am, edited 1 time in total.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Widebands That You Would or Wouldn't Buy

Post by Fred »

JaniM wrote:One candidate Prosport, LSU4.9
http://www.prosport.com.tw/
Added, no votes, yet.

Thanks for posting that, Hell_Fire, appreciated.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
toalan
Wideband Wizard
Posts: 241
Joined: Tue Jun 10, 2008 2:53 am
Location: Toronto Canada
Contact:

Re: Widebands That You Would or Wouldn't Buy

Post by toalan »

Fred wrote:I have a question, when you say "accurate" how accurate meats the definition of "accurate", when you say inaccurate you give a number eg 0.5 or 1.0 (:-o!!!), so is accurate <0.1 or <0.01 or?

May I interpret your post to place a "would buy" against your own, techedge, innovate and plx, a "would not" against the others you listed, and nothing against any others that are there (if any)?

Another question, when you say you value reliability, do you mean failure or do you mean consistency of results?

Personally I want it to be accurate and fast, and not fail, in that order. If it does fail, that's OK, as long as I know about it and can replace it with a superior device. Having it slow is a huge problem while trying to tune based on logs with any dynamic content. Have you ever driven anywhere without moving your right foot for significant periods of time? I doubt it :-)

With respect to accuracy, on the rich side, I'd want it to be wrong saying leaner than reality. Any amount richer than reality invites you to tune your car leaner than you think, and melt things. On that basis 0.5 AFR richer than reality is an utter utter fail for me.

FWIW, I only put current things in the list. If I was putting older stuff, your old JAW units would have got a negative vote. No favourites here either :-)

I really value and appreciate your input on this! Thanks :-)

Fred.
The units are say that are accurate is usually 0.01-0.02 lambda accurate with a new sensor.

I would put the Techedge units as a recommended buy, PLX and LC1 due to concerns of reliability I would not recommend. The rest I would put as not recommended. For the AEM and the NGK, I would say half not recommend as I only tested one unit and they are a big company so maybe I just got a dud.

My definition of reliability has mostly to do with the component quality and the voltage protection, in the case of LC1 it based on the countless "error 8" problems that everyone including myself encounters on a regular basis.

From my own tuning experience, I mostly care about WOT AFRs, I tune that first then I fiddle with part throttle response. If the wideband is slow, then part throttle maybe too rich/lean, but it is usually WOT lean that causes the biggest stress to the engines. If you take your time, you can do a good job tuning part throttle response even with a slow wideband. Now if you are doing real time closed loop engine control then I have no idea.

I would put my own JAW on the not recommend list, in my defense; it was my first wideband and my first electronic product, I sold it as a kit between 2007-2008, I sold it for $45 which at that time was half the price of other DIY lambda controllers.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Widebands That You Would or Wouldn't Buy

Post by Fred »

Re accuracy, wow, that's good! :-)

I understand the LC1 issues/reasoning, and was expecting some negative votes because of that.

I'm still a bit confused about how your votes sit. I see a definite positive for TechEdge and negative for all others, and no neutrals, is that right? Or neutral on ones you didn't mention? Or neutral on the two you're giving the benefit of the doubt?

About tuning and WOT, me too. In a high performance setup we're talking very vary rapid acceleration runs. At which point your entire lambda curve is lagging your engine's actual data, and becomes virtually meaningless in terms of correlating a value with a load point. The errors would just be too large. If you can steady state the engine at various load points, eg on a dyno, then it ceases to matter. But you can't reliably steady state an engine near the redline when it's spinning at 15k there. You simply must sweep up. When I tuned my truck (equipped with LC-1 for the time being, and way back then) I did it in the rain, with ~400hp, no traction, greasy roads, and I did it sideways. I would get on the gas and the RPM would shoot up in the blink of an eye. My passenger (who did the adjustments) was told to scream "stop" if he saw any lean at all, and then richen from there up. We got it pretty good doing that inside about 2 hours total over 2 nights. ~400hp spins a disconnected drivetrain to redline in the blink of an eye. A slow wideband would potentially miss the entire event. How slow are we talking, exactly?

No defense needed for JAW, I for one am glad you did what you did, the way you did it, learned what you learned, and now know what you know!

So I assume that your own SLC and Spartan units are also accurate to within the quoted range above? If not, you'd not be talking about others the way you do, I guess. And from that, am I to conclude that the only widebands on the market that check all three boxes (reliability, speed and accuracy), are yours and perhaps the new LC2? I guess it doesn't really surprise me if so :-)

And yeah, I know it's not your style to blow your own trumpet. Sorry to put you on the spot like that.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
baldur
QFP80 - Contributor
Posts: 44
Joined: Tue Aug 19, 2008 4:05 pm

Re: Widebands That You Would or Wouldn't Buy

Post by baldur »

I'm going to say would use the PLX, I know they use cheap components and I think the DC barrel connector is a joke but I've not had any trouble with them as long as I've managed to keep the power polarity right. I currently own two, both of which see regular use.
re Innovate, I have had trouble with calibration on early LC-1 units but firmware updates seem to have fixed those issues. I do have several LC-1 units and they're all 6-8 years old and haven't failed yet.
toalan
Wideband Wizard
Posts: 241
Joined: Tue Jun 10, 2008 2:53 am
Location: Toronto Canada
Contact:

Re: Widebands That You Would or Wouldn't Buy

Post by toalan »

Fred wrote:Re accuracy, wow, that's good! :-)

I understand the LC1 issues/reasoning, and was expecting some negative votes because of that.

I'm still a bit confused about how your votes sit. I see a definite positive for TechEdge and negative for all others, and no neutrals, is that right? Or neutral on ones you didn't mention? Or neutral on the two you're giving the benefit of the doubt?

About tuning and WOT, me too. In a high performance setup we're talking very vary rapid acceleration runs. At which point your entire lambda curve is lagging your engine's actual data, and becomes virtually meaningless in terms of correlating a value with a load point. The errors would just be too large. If you can steady state the engine at various load points, eg on a dyno, then it ceases to matter. But you can't reliably steady state an engine near the redline when it's spinning at 15k there. You simply must sweep up. When I tuned my truck (equipped with LC-1 for the time being, and way back then) I did it in the rain, with ~400hp, no traction, greasy roads, and I did it sideways. I would get on the gas and the RPM would shoot up in the blink of an eye. My passenger (who did the adjustments) was told to scream "stop" if he saw any lean at all, and then richen from there up. We got it pretty good doing that inside about 2 hours total over 2 nights. ~400hp spins a disconnected drivetrain to redline in the blink of an eye. A slow wideband would potentially miss the entire event. How slow are we talking, exactly?

No defense needed for JAW, I for one am glad you did what you did, the way you did it, learned what you learned, and now know what you know!

So I assume that your own SLC and Spartan units are also accurate to within the quoted range above? If not, you'd not be talking about others the way you do, I guess. And from that, am I to conclude that the only widebands on the market that check all three boxes (reliability, speed and accuracy), are yours and perhaps the new LC2? I guess it doesn't really surprise me if so :-)

And yeah, I know it's not your style to blow your own trumpet. Sorry to put you on the spot like that.

Fred.
I am not going to put any of them on the recommended list, for the do not recommend list I will put down
-BPSX
-NTK AFK
-AEM Controller in a gauge
-Prosport/Glowshift (they are the same company)

I found tuning to be an enjoyable labor of love, once you start tuning your daily beater you never really stop tuning it. So I could live with slowish response time as I was never in a hurry to get that perfect tune, over a period of month with some constant fiddling I eventually got that perfect tune I was looking for.

My SLC products will hit 0.01 lambda accuracy without calibration, spartan hits much better numbers than 0.01 lambda accuracy without calibration in practice but since the Bosch sensors are rated for 0.01 lambda accuracy I can not concretely claim any better than 0.01 lambda for spartan.

I am still working on SLC Free right now, I will start to put out kits in a month or so. I am pretty sure I can confidently hit 0.01 lambda accuracy with great response times on SLC Free. On a closed sourced product in which I can fully test the product and do calibration, it is not problem to hit 0.01 lambda accuracy, SLC Free is a significant challenge since I do not have a hand in testing and calibrating each unit. If I can prove that SLC Free is an accurate, reliable and a fast product even without my direct involvement, hopefully that gives me and my closed sourced products more credibility.
Post Reply