kiggly
Supporting Vendor
- 321
- 199
- Feb 17, 2003
-
Ann Arbor,
Michigan
Originally posted on Facebook, but obviously should also be here.
Warning: long post with a deep dive into squiggly lines.
To complement the upcoming auto trans straight cut planetary gears, we have been working on figuring out some perpetual shortcomings in the DSM auto trans. Today’s lesson is the 2-3 shift. This is the worst shift in this transmission. I made a transmission bench test using a hydraulic pump to create a nice, repeatable system. This is logged with my old Haltech Elite 2500 in ‘always on’ logging mode, which is perfect for tests like this.
When the 2-3 shift is commanded, the 2nd gear band/servo release is pressurized, the front/3rd gear clutch is applied, and then the end clutch is applied after these build some pressure. In OEM trim the transmission flow paths are very restrictive. Due to this, 2nd gear releases slowly and 3rd gear applies slowly, as shown in the red and pink traces. These clutches are on the exact same flow circuit (paired together in the valvebody). The circuit has reduced pressure until the 2nd gear servo fully strokes, which took about 2 seconds in the bench test. Third gear capacity is reduced until it has full pressure. The slow shift results in a lot of overlap between 2nd and 3rd gear, causing heat, wear, and binding in the planetary (extra torque). During the testing on Warren’s car, we could see the wheels visibly full stop on the lift during a 2-3 shift at 2000rpm. It is a mess.
The Translab shift kit improves the OEM behavior considerably, but it still does all the same stuff - just a little faster. In the Laser at ~210psi it took over a second for full pressure to happen in the 2-3 shift.
Luckily, there is major room for improvement. There is opportunity to dramatically increase flow path capacity in the valvebody with bypassing some ports and modifying OEM parts (details will be shared soon). Several iterations later, I have a behavior I’m pretty happy with. This setup removes the end clutch and uses its flow to feed the band release quicker. The front clutch is fed externally with a solenoid, using the rear clutch feed as its supply during 3rd gear. So, instead of using a single feed for 3rd gear, this is using 3 flow paths in the valvebody and also adding flow capacity via some bypasses. The front clutch feed timing can be controlled independently, so any binding or flare can be tuned out in electronics. The front clutch external solenoid controlled feed also functions as a trans brake.
Thanks,
Kevin
Warning: long post with a deep dive into squiggly lines.
To complement the upcoming auto trans straight cut planetary gears, we have been working on figuring out some perpetual shortcomings in the DSM auto trans. Today’s lesson is the 2-3 shift. This is the worst shift in this transmission. I made a transmission bench test using a hydraulic pump to create a nice, repeatable system. This is logged with my old Haltech Elite 2500 in ‘always on’ logging mode, which is perfect for tests like this.
When the 2-3 shift is commanded, the 2nd gear band/servo release is pressurized, the front/3rd gear clutch is applied, and then the end clutch is applied after these build some pressure. In OEM trim the transmission flow paths are very restrictive. Due to this, 2nd gear releases slowly and 3rd gear applies slowly, as shown in the red and pink traces. These clutches are on the exact same flow circuit (paired together in the valvebody). The circuit has reduced pressure until the 2nd gear servo fully strokes, which took about 2 seconds in the bench test. Third gear capacity is reduced until it has full pressure. The slow shift results in a lot of overlap between 2nd and 3rd gear, causing heat, wear, and binding in the planetary (extra torque). During the testing on Warren’s car, we could see the wheels visibly full stop on the lift during a 2-3 shift at 2000rpm. It is a mess.
The Translab shift kit improves the OEM behavior considerably, but it still does all the same stuff - just a little faster. In the Laser at ~210psi it took over a second for full pressure to happen in the 2-3 shift.
Luckily, there is major room for improvement. There is opportunity to dramatically increase flow path capacity in the valvebody with bypassing some ports and modifying OEM parts (details will be shared soon). Several iterations later, I have a behavior I’m pretty happy with. This setup removes the end clutch and uses its flow to feed the band release quicker. The front clutch is fed externally with a solenoid, using the rear clutch feed as its supply during 3rd gear. So, instead of using a single feed for 3rd gear, this is using 3 flow paths in the valvebody and also adding flow capacity via some bypasses. The front clutch feed timing can be controlled independently, so any binding or flare can be tuned out in electronics. The front clutch external solenoid controlled feed also functions as a trans brake.
Thanks,
Kevin
You must be logged in to view this image or video.
