Need differential signal to auto steer

I am using an RS+ as GPS source for a tractor steering guidance field computer. The computer/display console has an internal GPS with WAAS that has been deselected to permit the application of this external source. I purchased a single RS+ for proof of concept that it would actually talk to this steering system. BTW, many thanks to PotatoFarmer’s contributions to the forum in making that happen. I have hopes of adding another as base to achieve economy RTK.

This arrangement works for making manual steering inputs using the console display for guidance. The problem is, with the RS+ connected, the warning “No Differential” is displayed and the auto steering function will not engage.

According to the manufacturer’s tech support, the auto steering node, to achieve acceptable accuracy, is designed to only accept GPS data having a source of augmentation. The “No Differential” warning indicates that WAAS corrections are not being received. Upon researching the forums here I was disappointed to discover that the RS+ does not support WAAS.

Is there a workaround for this? If corrections from another RS+ as base station were sent to the rover, would that be forwarded to the steering computer as a differential correction message? Or is correction only between the Reach units and rover is sending the same data to steering, it just happens to be more accurate?

Edit: Forgive me if I have mixed or misused terminology. I am not certain as to the differences in augmentation, corrections and differential.

Thanks

If you have just one Receiver RS+ you have to use NTRIP option (you will need internet connection of course) normally you obtain a good accuracy.

https://docs.emlid.com/reachrs/common/quickstart/ntrip-workflow/

Yes, I understand that would be another way to achieve the highest possible accuracy.

To rephrase my question; The steering computer is expecting DGPS data, is there a NMEA 0183 sentence generated by a DGPS receiver that identifies itself as capable of providing a differentially corrected position? Be it WAAS, RTCM or RTK from another RS+ as base?

Honestly I’m not sure, I’ve never tried it with EMLID receivers.
But one thing is certain is that RS+ allows to output NMEA format through blootooth for example.
Maybe a specialist could answer you better.

Thanks anyway for giving it some thought.

After taking a look at some of the GGA sentences, it looks like the fix quality is mostly 1’s for GPS, rather than 2’s for DGPS. For the sentences I looked at, it was using at least ten satellites.

Is there an expert available that can tell me if a WAAS enabled receiver, given a view of 10 satellites, would report that as DGPS fix quality? And if a receiver was not using any SBAS corrections, but was taking correction from a local base unit, what would the GGA sentence have in the fix quality entry?

Ok not sure of your exact issue or type of autosteer you are integrating to, but here is the common rundown for ag steering systems rejecting the mighty powers of emlid.

  1. Designed to use the talker code “GP” emlid uses gnss so its talker code is “GN” so rejecting GNGGA GNVTG, vs GPGGA GNVTG

  2. Too much data, some are only looking only for one or two pieces of data. If they have extra data that is not needed, it dumps the whole string.

These two issues are now on Emlids to do list for some future date!

  1. If its like my good buddy trimble 150’s older brothers; 250, 500, 750… they will only do field mapping only with a generic NMEA input. You need to get past an arbitrary software protocol called TSIP to do guidance. The 150 was obsolete long before these awesome little emlids existed, so they forgot to lock that up. Also so it could still work with the old 252 dinosaur base station.

4.RTK is a form differential gps but DGPS is not RTK.
DGPS settings in an ag guidance system usually refer to the Australian DGPS system, cause they were way cooler way earlier than the rest of us.

To get an Emlid to work on a guidance system you only want the Emlid to send position to the autosteer computer. But it must be one that can understand the rtk sentences. The 150 displays fix, float, or no corrections depending on whats happening to the emlid.

  1. If you are indoors and the emlid cant get to float and it usually will not start putting out NMEA data.

All corrections are handled Emlid side, my autosteer will not work until i get a fix. I use 2 M2’s. If the m2 does not get corrections, but gets to float, the trimble flashes no corrections.

I know these little emlids are still pricy, but they are still far better than a trip to john deere for an unlock and subscription lol. They really do work better in pairs for rtk.

  1. An emlid usually will not work as base for a Autosteer even with rtk corrections unlocked. They still use RTCM3.x(0) as the standard setting not 3.2. So you basically have to shell out the $20k dollars for a basestation due to that little issue or pay the yearly ransom for each tractor, or really madly deeply build your own translator and it takes forever.

It took me two years to generate my first correction stream to work with an rtk unlocked 750. Still only picks up 8 gps sats. Have to pay again to unlock gnss.

It took me two months to figure out how to integrate emlid to the 150. It now has more features and better reception than the 750.

The more data you keep handled by the emlid the better the system.

Not sure how much that helps, but i did a lot of research here on this forum that helped me greatly.

Hopefully you can get it to work, the first time I saw the trimble say fix i was so elated! What a great feeling solving difficult puzzles!
Keep trying, read manuals, you can do it!

3 Likes

PotatoFarmer, thank you for your reply! I have been scheming about a way to get cheaper and simpler precision guidance into my tractor for a long time. It wasn’t until you posted your project and revealed the need to change talker ID’s that I went ahead and bought an RS+ and a Shipmodul multiplexer.

I am working with a Raven Cruiser II connected to a Smartrax MD steering computer. Raven offers a cable that is used to interface their own modem, hub and globe for RTK guidance. It only amounts to a DB9 male plug labeled DGPS Input with pins 2,3 and 5. These correspond with the DB9 pins on the Emlid cable.

Having pins 2 and 3 (and 3 and 2) measuring -5v and 0v in reference to pin 5, I surmised that a DTE/DCE RS232 relationship must exist. Using your idea of changing talker ID and matching baud rates I was able to get NMEA 0183 into the steering node and on to the field computer console. This console offers the option to deselect it’s internal WAAS GPS and use the external input.

The console then would display AB line guidance with EMLID input, but probably at less precision without WAAS. It didn’t seem to care that it wasn’t differentially augmented at that level. But it did generate a warning message on the screen saying No Differential, and the auto steering wouldn’t engage.

I talked to a Raven tech who explained that it wouldn’t engage because it wasn’t receiving WAAS corrected position data. The question remains, is it dissatisfied for lack of WAAS, or lack of any type of differential correction? The Smartrax system is advertised to be RTK capable right out of the box.

This led me to find out where in the NMEA data this distinction is made. It seems to be in the GGA sentence at field 7 called Fix Quality, a single digit number following W (or E for the other hemisphere) of longitude. The range is 0 to 9. Reading the Emlid Reach GGA output on the configuration screen of my Shipmodul MiniPlex, it is always a 1. This corresponds to “Single point”, that makes sense. Looking at the forum photo of your MiniPlex output, the GGA sentence shows a 4 in that field, which corresponds to “RTK fixed ambiguity solution”, that makes sense too.

It could be that the steering computer is looking for only a number that corresponds with WAAS Fix Quality, which is the number 9. At this point it looks like my only option is to buy another RS+ and set up a base/rover configuration and feed it a number 4 to see what it thinks of that.

Thanks again for your contribution to bringing precision to the little guy.

3 Likes

Thank you, I am excited that someone else is trying this solution.

I think you are on the right path. The 150 will not engage auto steer guidance until the emlid gives the “4” as you described.

It will maintain auto steer if it falls back from fix to float for a decent amount of time.

You will be shocked how straight your lines become with RTK. Emlid really made a good product for tractor guidance and it is always getting better.

2 Likes

Hi everyone,

Thank you all for your discussion! I’ve learned some new things from this thread :slightly_smiling_face:

I just want to comment on the WAAS question. We indeed don’t have WAAS support, and so far, there are no plans. WAAS is good for some purposes. However, we’re mostly focusing on getting the centimeter-level accuracy and improving RTK performance for our receivers.

1 Like

Saying I was disappointed WAAS was not supported was too strong a word. Surprised would be more accurate. But as I think about it, WAAS is not doing me any good anyway, otherwise I would not be here seeking centimeter level accuracy!

I am looking forward to my second RS+ arriving this week. I will report back on how the auto steering integration works out.

3 Likes

Once you have your own basestation you will forget WAAS ever happened.

Your next quest will be getting your corrections to transmit the maximum distance. Lora can do a section of land pretty easy if its at its base antenna is at the highest point.

Lora is pretty slick. Right now I pulled my rtk box apart to wire in a second comm port to accept corrections from higher powered serial radios. Hoping for better distance, and to experiment with higher correction rates. Though 1Hz corrections seem to work just fine.

Cellular is not hard to do either but then you have the extra bill to pay all the time.

1 Like

I have not done it yet but, on a Trimble base station we use a 900 Mhz 10 amp inline amplifier. Boosted our range considerably for about $450 US. I am sure it would work with the Lora radio just need a SMA to Type N connector and it would be up and running.

BTW, this is on a 90 foot tower as well.

I am hoping to get at least a mile radius of coverage from a single base position. I don’t mind having to move the base for the infrequent need beyond that.

Plus I only need relative accuracy, and plan on always using the same manual input position for the base. Am I correct in expecting this will provide repeatability? I am hoping to plant within a few inches of the previous year’s rows. It would be acceptable to nudge the line laterally, but can I expect this years AB line to be parallel to last year’s?

I should clarify, the AB line will be taken from the previous year”s job record so it would have the same lat/lon attributes.

I understand that the unit can be powered up remotely through cable. Is there an RS+ user that could tell me if this unit can be powered down remotely?

The emlid 0.1w lora should do a mile radius with ease at 1hz corrections as long as you can see the antenna. Try to put it up high on a bin, shop, pole, hill or tower. Trees and hills kill the signal pretty quickly, especially hills.

As for repeatability you are in for a treat. If you just need pass to pass to match year after year its simple. RTK is king of repeatability.

  1. Mount your antenna on a fixed location so if it needs replacement at anytime you know the exact spot.

  2. Let the Emlid take an average float to get base coordinates on the base tab. I like to let it take the full 30min but for pass to pass any number it comes up with will do.

  3. Save the time weighted base coordinates to manual. Set your antenna height, there is a max I think its 4m? If you are higher than this number just put the max value.

  4. Take a screenshot of your settings so you never forget them.

  5. As long as the base antenna is in that exact spot, with those specific coordinates entered, your AB line will remain in the same spot year after year after year.

Your base for pass to pass guidance does not need absolute geodetic position accuracy like the Surveyors need.

Think of the RTK base like laying local gridlines. The gridlines are accurate to each other regardless if the relative base coordinate is offset from the absolute location. But if your base antenna was to move over one foot for some reason, if the entered coordinates have not changed, your grid will be shifted over one foot too, so will your tractor.

Read Emlids doc about setting up a base it explains this very all well.

RTK is wild with the zero drift.

With WAAS i could barely have a sandwich, or finish a row and it would be drifting a few feet over.

With RTK a month later and I am still in the same tracks. Next spring will be in the same tracks too.

Here is a friend of mine swathing with RTK. We set up his own cellular base. He did a bunch of guess rows just cause he could and did not leave any strips!

1 Like

Awesome stuff

I have acquired a second RS+ and set it up as a base unit to provide RTK fixed corrections to my tractor auto steer. The steering node will accept them after they pass through a multiplexer to select only the necessary sentences and have the talker ID’s reassigned as all GPS. While this is an encouraging breakthrough, the results are less than optimal.

When capturing and maintaining an AB line it is slow to begin corrections, as a result they are abrupt and excessive. It seemed to me that the steering node was not getting info fast enough to keep up. I adjusted the steering computer’s sensitivity settings both up and down. Some improvement was found with an increase in aggressiveness.

It also became more responsive after increasing baud rate to the steering node to 57600. When using the field computer’s integral 10hz WAAS gps the steering worked well at 19200. I also increased the base update rate from 1hz to 5hz. Not sure if that helped.

It doesn’t make sense to me that the system behavior should change with a different source of positioning data. The line it is following is not any straighter, it is only more consistently defined. The WAAS unit did a great job of acquiring and tracking a straight line, the only problem was the pass to pass drift of that line.

For those here that are auto steering something, have you had good results with the update rates recommended?

I can understand why they would be suited to the ground speed of surveying applications, but at 5 to 10 mph are they fast enough?

Could LoRa air data rates or rtcm3 rates have any influence on responsiveness of hardware?

PotatoFarmer, I would be interested in what type of auto steering you are using and how you get it to track smoothly.

Allright there are three separations to make and tune.

  1. GPS/GNSS reciever / Shipmodul
  2. Guidance Parameters and Computation
  3. Limitations of your Equipment / (steering play, hydraulic speed, ground conditions, Implement drag.)

1 The tuning of the Emlid is not tough, you do have RS+ they work good but there is a reason that the RS2 came out. RS+ is only L1 GNSS, it takes longer to get and hold a fix compared to the RS2. When i was working with the M+ I could sit around for 5 or 10min to get a fix, and if lost the fix would take some time to regain. Then M2 is L1/L2/L5 and fixes almost instantly once warmed up.

If your Guidance can take 10hz Set all of the RTK rates in the Rover to 10hz that you can use. I am in Northern Alberta so i leave QZSS unchecked as i will never see it high enough on the horizon to be useful. Leave any setting you are not using off or unchecked to save processing power.

I found a quirk with the shipmodul, if you are inputting data to it at 57600bps for example and the unused inputs are set at 4800 all of a sudden it is using high buffer. set all inputs even if not used to 57600bps to match the used input. Then no buffering occurs. click the box “real time data” on the input you are using. Make sure your output filter is only sending the sentences you need to the output. I also found some added success recently changing to a ttl to 422 converter for the Emlid to shipmodul input. Though they let you creatively wire in 232, it is much more solid at the higher input voltage of 422.

The shipmodul may be the weak link. I am only running 5hz, but the emlid NMEA dumps so many sentences into it that i am sure the shipmodul is having fun getting all the data sorted out at 57600 which is high as mine can go. Once Emlid allows NMEA sentence selection and talker code transformation we can drop these multiplexers all together, and probably see yet more steering improvement.

Try some sort of vibration isolation on the RS+ Rover, the GNSS is so precise it ends up calculating vibration as position. Reducing any amount of vibration will reduce the calculation load on the processor.

Base settings for Alberta are GPS+GLONASS+GALILEO+BEIDOU 1HZ , Plus ARP 0.1Hz, Lora set to max data rate and max db.
The Emlid Loras will clip and drop connection with corrections set higher than 1hz on the M2.

Your Rover Setting would be GPS+GLONASS+GALILEO+BEIDOU 10 HZ
This automatically sets your NMEA to 10Hz

In the shipmodul make sure it is only transmitting out the needed sentences, for example the 150 only needs GPGGA and GPVTG.

If its only two or three sentences 38400 is requested in the manual at 5hz by Trimble, if it was at 10hz I would expect to be running at 57600 or 115200. The idea is to get the data in as fast as possible so it can start its position calculation and finish before you have moved.

  1. Your WAAS tuning will not work for RTK, this is expected.

I would start from scratch and precision measure your equipment again. This is the #1 reason on all guidance platforms that I have fixed that were not working properly. If the tractors measurements are garbage it will control like garbage. The biggest error every time is the wheelbase of an articulated 4wd Tractor, they measure only one side and get it very wrong. Measure front of rim to front of rim on both sides, Add those together and divide by 2. Now you have the correct wheelbase, all of a sudden their steering improves.

Tuning to WAAS is like trying to shoot a duck that cannot figure out which way it wants to fly. The position input hops back and forth so badly that you are hoping that half the time you are steering to the middle. RTK is precise so now you have to tune to an actual position, it really changes the dynamics of the tuning. But once you have it nailed down to cm steering, it never changes.

Watch your correction age it should be 1-2 seconds old max, if it is going higher you have radio setting issues.

I am not sure how a Raven tunes. But the Trimble for example my steering angle went from 12 to 18, WAAS vs RTK. I finally found my steering free play settings were far greater than I first imagined on a tractor with under 1000 hours. Aggressiveness is the very last thing to tune to field conditions.

My RTK tuning steps for Trimble 150 are:

  1. Set aggressiveness to 120
  2. Set steering motor to max speed
  3. Set steering free play to a minimum of 2.5"
  4. Decrease the steering angle until you start oscillating, then add 2 or 3 Degrees more.
  5. If the Tractor is running to the left of the line constantly, you have to get it to turn right. Slowly add more to Right free play until tractor adjusts toward the line. Same goes for left. Left and right free play are rarely the same number. Free Play is the most important setting of them all.
  6. Lastly increase aggressiveness till you notice it is hunting back and forth. Then slowly reduce until you are running a smooth straight line.

If the raven has auto tune use the auto tune wizard with the RTK, it will come up with far different numbers than WAAS.

  1. Depending on age, wear of components, speed of hydraulics, other oddities you may not be able to achieve perfect steering but pretty close to it.

I find that if you are using a steering wheel motor, rubberizing the foam roller really helps on slippery Deere steering wheels. I use a couple wraps of rubber tape just tight enough to stay on. The added traction really improves performance. Rubber tape is also called splice tape and has no adhesive it only sticks to itself, used for waterproofing splices. It is not Electrical tape, that goes over splice tape.

Hopefully this helps you out.

I will be updating the Trimble 150 Conversion thread pretty soon. I have added the ability to hook up to an external 232 corrections radio. Still testing lol. :grin:

2 Likes

PotatoFarmer, I just skimmed over this and it looks like plenty of good information to work with. I am going to the tractor tomorrow and will dissect it completely, applying what I can. I will likely have even more questions then.

Thanks for the quick response and sharing your hard earned knowledge!

1 Like

A question already. I originally had a 232 to 422 converter between Emlid and Miniplex3. When I was having no luck getting connected I contacted Shipmodul for support. It turned out the issue was not the converter, but I was told it wasn’t necessary anyway so I don’t use it.

You mention a ttl to 422 converter helped your setup. Is my converter similar, and should I not send it back to the supplier?
I actually have two, planning on needing one reversed on Miniplex3 output to steering node, but was told by Shipmodul not use that either.