Jump to content


Full Members
  • Content Count

  • Joined

  • Last visited

Everything posted by trevmar

  1. @MRN76 Great to hear from you again! With now an Android rollback app! Impressive! Was there some reason the message system would not deliver messages to you last week?
  2. There was a charger sold by More4Mini which gives a 64V charge rather than 60V. This gives about 15% more capacity, but at the expense of battery lifetime. The normal 60V charger only sends each LiPo to 4.0V, rather than 4.2 Inside the battery pack there is a battery management chip which isolates the charging for each cell. When the battery-pack senses it is in distress you have to get past that chip and trickle-charge the worst cells. Different techniques are needed for different levels of failure.
  3. @FreeRide Generally I agree with you. The firmware is probably v1.1.1 as it reported, but perhaps with the pushback having been turned down a little. None of us seem to have any firmware before 1.1.7 to compare it with, and there is a HUGE difference between v1.1.7 and the "v1.1.1" in the pushback . However, the fact that they tried to protect the CPU firmware by locking it with RDP1, the same security used on Swallowbot, indicates to me that somebody did something valuable that they wanted to protect. So it is hard to draw solid conclusions. What amazes me is how much 'better' the Chinese Ripoff-bot feels to ride, yet it seems equally safe when the miniPRO is being stressed. To me, it emphasizes how much Ninebot damaged the utility of their platform in their futile quest for "safety." The v1.30, for example, and Swallowbot itself, is a pain to ride uphill because it keeps incrementally pushing back the platform as you get close to its lifting limit. Even 1.1.7, which doesn't have this behavior, pushes back so aggressively when its limit is reached that it very nearly pushes even an experienced rider off on the steepest slopes. The original design team, those who produced v1.1.1, did a much better job than those trusted with the machine during the marketing phase. @WEagle95 was correct in his warning (in a post on June 27, 2018) about deficiencies in "critical IRQ handlers .. what we must do on a "hard fault" or "watch dog" interrupts in mission critical applications" (I consider my riding it as "mission critical" LOL). He tells us Ninebot just switches the platform off rather than trying to go to an "emergency mode." I have confirmed this. I have been looking over the various firmware code versions for some time now. @WEagle95 really did a great job with Swallowbot (with the tilting platform exception). Yet this was achieved without only two small routines added to the original v1.3.0 binary code, and quite a lot of word-level patching of the various behaviors of the v1.3.0 code. As for Ninebot shutting reverse engineering down, in the USA you are allowed to repair a device you bought (which we now know would be to v1.1.1 presumably). The DMCA Trennial Review made that quite clear ( https://s3.amazonaws.com/public-inspection.federalregister.gov/2018-23241.pdf ) specifically "Computer programs that are contained in and control the functioning of a lawfully acquired motorized land vehicle" and I look forward to Ninebot arguing their way around that one... For those that missed it, I have attached a screenshot of @WEagle95 's last post on his Swallowbot work:
  4. Jojo - It looks like all the Ebay and Amazon sellers have removed their Retro-bot control boards from sale in the last few hours from both Ebay and Amazon. This is a pity, as I am very happy with the v1.1.1 firmware they were shipping. The Russian Swallowbot seems to have gone defunct too - I can't send a message to @MRN76 any more, and @WEagle95 hasn't posted for months.
  5. It looks like all the Ebay and Amazon sellers have removed their control baords from sale in the last few hours from both Ebay and Amazon. This is a pity, as I am very happy with the firmware they were shipping. Even if it is not the Russian Swallowbot (which seems to have gone defunct too - I cant send a PM to @MRN76 any more, and @WEagle95 hasn't posted for months). Consequently I started a new topic: https://forum.electricunicycle.org/topic/14057-sharing-experience-with-russian-swallowbot/
  6. Jojo, I have an old version of the app, (v4.4.6), so the balance slider goes from 0 to 5, with 3.2 recommended. I set it to 5 (maximum) and put my feet further forward on the machine. Although I feel a little more secure, the "rolling" of the platform's balance point as I get near to maximum load is very unsettling. It seems as though Alex&Alex didn't change that function of firmware v1,3 at all. They did reduce the aggressive pushback at maximum load, however (for me, at about 10Kph uphill on our 12 degree grades).. No wonder these users who don't have 1.1.7 (or similar) firmware available are complaining so bitterly about their machines (running 1.3/1.4)! I did achieve a significant change in downhill performance by enabling the Speed Limit button for a few seconds. It looks as though that is something which needs to be done as 'initialization' of the Swallowbot installation. I have already changed from my "manual ride smoothness" to "automatic" and that seemed to change a little. I left the steering sensitivity on 'Automatic'. ps: This v4.4.6 version of the app is the only one I have ever got going, and it is in a spare Nexus-5X without cell service. All my active phones are Rooted and secured, and I haven't been able to get the Segway app to install in them pps: I am starting to become quite firm with my theory that the v1.1.1 Chinese 'Ripoff-bot' firmware may indeed be the firmware initially signed off by the initial development group, before marketing were given the chance to cripple the machine with speed limits, etc... If the rumor you heard was true, that the initial firmware was written by one developer in his school holidays (which is feasible, IMO) then the Ripoff-bot firmware will have to be renamed to something like 'Retrobot.' I am happy with that Chinese firmware, incidentally, so maybe we should start coming up with some better names for it pps: DId you ever see any firmware released from Ninebot lower than 1.1.7?
  7. I have a question for those of you who use a copy of Swallowbot -- What do you think of it? I finally got the Swallowbot CPU from MRN76 working. It announces itself to the app as "v7.6.0"? Yet I see some of the YouTube videos with different version numbers? Any idea why the difference? (AFAIK, WEagle95 only ever posted notice of one development release). The problem I have with it stems from the underlying Ninebot 1.3.0 firmware. I hated this firmware with passion - and changed all my machines back to 1.1.7 (using the JTAG method and the software from MRN76). What I hate is the progressive tilt-back. As you get towards the load limit the feet are incrementally tilted back, little by little, and if you struggle against this to accelerate you get to the point where the platform tilts back violently with beeping, at that point pretty close to tipping you right off. The older firmware versions (up to at least 1.1.9) do not do that. Whenever I go uphill, Swallowbot does this too. Swallowbot does suppress the final violent attempt to throw you off while beeping, but the slowly tilting back platform during uphill acceleration is just the same, and just as annoying. Going on flat surfaces, or downhill, Swallowbot is effortless, and a pleasure to ride.Has anybody figured out a way to manage the uphill grades? Interestingly, the Chinese 'ripoff-bot' firmware does not do this. It announces itself as v1.1.1, and does indeed seem based on this initial firmware. I haven't been able to test its maximum speed, but it is extremely gentle when it tilts after reaching load limit uphill. A pleasure to ride (if you aren't bothered by all the beeping). It seems to have the same high performance at lower battery levels as the Swallowbot. Currently I have Swallowbot in one miniPRO, Ripoff-Bot in another, and v1.1.7 in the miniPRO with knobbly tyres. Each has benefits and each its disadvantages. What have I forgotten??
  8. I used a gear-puller.I left one of the two bolts in, but loose, so I had something to push against Like this:
  9. @FreeRide I have been riding around two identical miniPRO, one with a 310Wh battery and one with a the smaller 240Wh battery. Both batteries have matched cells, and are in good condition. But they each use different quality/chemistry LG-brand cells (see my topic on pulse testing of the batteries) and I get the feeling the cells' discharge characteristics are different. I feel that the cheaper cells in the 240Wh battery drop to 2 bars rather quickly and then stay at that level for much longer than I would expect. I think this is a function of poor BMS firmware, not the control board firmware. Has anybody else observed this?
  10. @FreeRide I found yesterday that my 2.4.6 app speedometer reaches half scale (12 o'clock) while showing about 3.8mph on Ninebot v1.1.7 and 3.5mph on the Chinese 'Swallowbot.' I do not believe those values, and will use a GPS app to check them today. The v1.1.7 machine does have larger knobbly tires. I am pleased with the way the Chinese firmware handles low battery, however. Indeed I am still using my test platform with the smaller 240W battery, routinely riding comfortably at 2 bars. I also am getting used to the gentle pushback to the point that I notice the beeping more than I notice pushback now. I have only seen these warnings when going uphill at a brisk pace, never on the flat or downhill. So I am getting more and more happy with the firmware, it seems to match my riding very well.
  11. @FreeRide It is pretty easy to flash in a Ninebot version of the firmware if you want to do that. I use an SWD programmer, with the Windows software developed by @MRN76 I am still testing to see what I think of the Chinese firmware, and, as I said, I am going to try the Russian Swallowbot CPU at some point to directly compare Chinese with Russian firmware mods. I never expected this Chinese firmware to be same as @MRN76 and @WEagle95 developed. You will recall @AndyYellow here who posted this about the Chinese Firmware he bought from TaoBao: https://forum.electricunicycle.org/topic/7293-hacking-the-minipros-speed-limiter/page/4/?tab=comments#comment-197016 So I have never assumed that the Chinese 'Swallowbot' was the same as the real Swallowbot from @MRN76 and @WEagle95 . However, I failed at getting their CPU working on my first attempt, and no other opportunity has arisen to get the Russian firmware. So I am looking to get the best MiniPRO experience I can. For me, the controlled pushback is a big deal. I also want it to be at least as sure as the Ninebot firmware when negotiating ledges and obstacles. I really am not as worried about the maximum speed as I am about getting a reliable, safe, platform. Finally, have you thought of the possibility that there is a 50Km 'bonus level' in this firmware, just as there is in the Ninebot firmware? And yes, I have considered that this firmware may just be the v1.1.1 Ninebot release. But Ninebot don't protect their firmware against copying, and the CPU I received has been secured with RDP1. So I am keeping an open mind, and becoming more comfortable as every mile goes on the odometer
  12. Ah, I see your screenshot. My 4.4.6 app has everything in it except the "Limited 17Km/h" text. I just double-checked. Never mind, I will look at its speed in an hour or two...
  13. @FreeRide My app (v4.4.6) doesn't tell me speed limits. But it has a speedometer with a 12-o'clock position that presumably is half maximum speed. I will check that out. Please let me know if there is any way to find in the app the maximum speed, I don't see it. Just the speedo. But I can run it to 50% and see what it says. I will check that out. What I did today was try to understand why my miniPRO uses 90 watts of energy to climb the 1-mile-long steep incline here. With my weight (95KG), running right at the point the 'swallowbot' starts beeping me all the way up the slope it takes about 90watts out of the battery, and with the smaller 240 battery I have fitted to the test-unit-miniPRO that drops me to 3 bars. A little more riding and I have dropped to 2 bars. Even at that level the miniPRO is nice to ride, comfortable, firm, a pleasure to ride. So that is good. However the incline cannot be taken faster in Swallowbot than with 1.1.7 firmware. So I must be right at the climbing limit, and that seems the same in the Chinese 'Swallowbot' as 1.1.7 firmware. Interestingly, I was watching this video of Russian Swallowbot and I noticed that it was giving the quick beeps that the Chinese firmware gives. I had thought that the Russian Swallowbot had a lower pitched warning tone, but I guess I was wrong. So the warning beeps are the same, and the pushback is very mild indeed. It is a pleasure to ride, even at low battery. https://www.youtube.com/watch?v=v8ivTNxFWUA
  14. Tommy, I am pretty sure the boards don't use the Russian Swallowbot firmware, although I find it strange that neither @MRN76 or @WEagle95 have confirmed or denied any collaboration. I am not sure that I needed to go through the break-in mile of riding, as I didn't try to click the "Limted" button until after I had done that. I did have to go through the rubbish tutorial about how to get on the machine, etc... I cannot yet confirm the presence of the 50Km threshold.. The board can use my 4.4.6 Ninebot app, but does not respond to the "Riding Sensitivity" or initial slope commands, as it does appear to be really based on v1.1.1 firmware. I have been slow in testing the maximum speeds as the battery in that machine has been misbehaving, and I have not seen speed extensions yet. The only confirmation that the firmware is different is that when I strained it too much going up the hill the push-back was far more restrained than any of the firmware versions I have previously ridden. Indeed, it seemed to be a good compromise between safety and convenience. But I have never ridden unmodified v1.1.1 before, as @MRN76 does not include that version in his "Ninebot.rar" firmware utility, and it is possible that this was its normal behavior. The Chinese CPU is protected from reading (copying) just as the Russian Swallowbot CPU was. Now that I have good control boards which are easy to solder, I will at some point test the Russian Swallowbot CPU from the board which previously blew up (due to the plastic coating on that board making my soldering fail). These Chinese boards are clean. and tempting me to solder the Russian chip again and see if it was damaged when the output FETs blew up. But first I will finish playing with the Chinese mods.
  15. The 320 indicates a bigger battery, which is what you should really be happy to see. Let's hope it was maintained in good condition at the warehouse (white box). The "2018 MiniPRO" sold on Amazon only had a 240W battery. So you struck it lucky. Probably a factory refurb, I bought 2 of them, although they came in normal boxes.
  16. Joe, added 4 July 2019. I was dropped by one of my miniPRO in exactly this way last week, and dislocated my shoulder. I think it was because I had the smaller 240Wh battery pack in it instead of the normal 310Wh I normally use. I was running some tests at lower battery powers. Battery status was 56% after it dropped me. Will update as I get a better idea of what happened. Test from 23 May 2019: Joe, We have hills around here with a 1:12 grade, lots of them. I have never been dropped going downhill on any of my MiniPRO, with widely varying firmware. Maybe your battery is in poor condition? Have you tried to find another nearby MiniPRO owner whom you could swap batteries with? Myself, I rarely drive immediately after charging, Also, has your charger 59 volts output with nothing plugged in? At one time they were selling 64V chargers which give a longer range,, but shorter battery life, and might perhaps trigger an over-voltage condition downhill, although I have never seen it here.
  17. I have connected the second Chinese 'Swallowbot' control board to a JTAG / SWD programmer and verified (with OpenOCD) that the CPU Firmware has been protected against reading to RDP1 level, as was the Russian chip I got from @MRN76. Actually that is somewhat good news, as Ninebot ships their normal CPU's unprotected. It looks like these CPUs have been soldered in by the seller, and that the contents of the CPU are protected against copying, as was the Russian Swallowbot CPU chip. So maybe the chip has something valuable in it. We shall see... Protection is inconvenient, but (I guess) expected, We will just have to determine the performance of the Chinese firmware by testing it on the road...
  18. The top speed with v1.1.7 is about the same with version 1.4.1 but the pushback is much less aggressive with the earlier versions of firmware. Just follow @MRN76's instructions at: http://mrn76.ru/index.php/programming (use CHROME browser or Google Translate to translate the Russian) Here is the link it gives to the Ninebot Firmware changing program: https://yadi.sk/d/C38vGNAI3LMqKq Everything was easy and straightforward and works well. I haven't been able to get a copy of SwallowBot, I 'broke' the chip that @MRN76 sold me, so currently I am working on trying to figure out what these Chinese sellers on Ebay are selling and calling "Swallowbot." I am posting my tests on that firmware in the normal forum.
  19. I want to get some confidence in the platform first. So I will be testing it out against the standard 1.1.7 where I know it fails, on the steep slope and at obstacles. You guys can all test its maximum speed. I have already ruined my ankle once, thank you... LOL
  20. The second board has arrived, this is the one I ordered a month ago from Ebay seller Harhap but which they didn't ship until last week. It is exactly the same as the first. brand new, no sign of thermal grease or bent wires to indicate it has ever been connected in a MiniPro chassis. Like the first, the CPU soldering is by hand, unlike the rest of the board. I haven't checked if these CPU are read-protected, but I will be surprised if they are not. Tomorrow I will get the chance to go up the hill to the 1:12 grade that the MiniPro has such trouble with. Let's see how the Chinese firmware handles that...
  21. Personally, I would swap both motors at once, just in case there was a significantly different torque between wheels due to differing manufacturing batch, and different strengths of the magnets they have in them.
  22. I forgot to mention that the older/hacked v1.1.1 Chinese firmware is working with the Segway/Ninebot app v4.4.6, which is a newer app version. Jojo reminded me that the original control boards with the original v1.1.1 do not work with these newer versions of the app, so I am gaining more confidence that the firmware is indeed a hacked version as it was advertised to be...
  23. Well, that's exactly what I was trying to find out... It is clear there are huge differences, but I am looking for somebody familiar with the Russian firmware to suggest some things I can easily test for (without risking my neck). I suspect at this point, however, that this is home-grown Chinese firmware. It may not even be enhanced from the old Ninebot v1.1.1, I haven't been able to determine that yet. My wild guess is that the original firmware team has taken out limits imposed on them by marketing and released this firmware from their archives. This guess is based on the version number being 1.1.1, and my suspicion, based on years of heading firmware projects, that the original team was thrown out once they got the product solid enough for the market. It is all too common for marketeers to come in and 'restructure' the engineering team with "yes men" rather than the innovators who gave the product life..
  24. Are those the values from Alexander for the Russian Swallowbot or for these Chinese control boards?
  25. The first control board, the one I ordered last week from the Amazon marketplace seller EKSmart, has arrived by DHL from Hong Kong. It doesn't look exactly like MRN96's Swallowbot, it reports firmware v1.1.1 and serial number 11960/100112*** (last 3 digits blanked for security). It made me go through the "learning exercises" when I switched the machine on. I suppose the good news is that the board works at all! I have ridden it around a little, and it didn't disable the "limited" mode automatically, I had to do that via the "Limited" button on the app. Good news is that I have to push it hard to hit any limits, bad news is that it did push me back in the "limited" mode. But not so much. There was noticeable tilting of the feet in limited mode, too - trying to give me feedback, I think. That has not yet been obvious in unlimited mode. The alarm beep is a very quick sequence of about 6 normal-pitch beeps. Maybe the Swallowbot owners can give me some idea of how to test it without trying to go past 18Km/h After I have ridden it a bit I will work up to this, but since this is my first ride after the winter break, I didn't want to push things too hard until I get confidence in the firmware. I still am still waiting for the board from the Ebay seller which I ordered and paid express shipping on a month ago, that should get here Monday, I think. The month delay was clearly to get new stock - it seems the first set of boards sold out from all the Ebay sellers. I want to emphasize that the sellers are all calling the firmware "Swallowbot" even if it turns out to be something quite different from what @WEagle95 and @MRN76 developed. ps: I checked the SKU number on the packaging SKU01399 and found hits on Alibaba saying it "Brand: Cool Step", customized with pre-installed firmware swallowbot, for mini-pro,mini-lite
  • Create New...