On 07/02/2011 03:44 PM, boB wrote:> On Sat, 2 Jul 2011 12:24:56 -0700 (PDT), Jerry Avins<jya@ieee.org> > wrote: > >> On Saturday, July 2, 2011 3:16:19 PM UTC-4, tinkerz wrote: >>> I am after a control system based on frequencies, I need to gather >>> information on the time and frequency domain. >>> >>> The data is fast and noisy, I made an assumption the fourier and laplace >>> will be the place to start. >> >> You need to lay aside your preconceptions and find a good consultant. As a former consultant, I felt it was my duty to give the clients what they needed. That was often at odds with what they wanted. Tact usually won the day. >> >> Jerry > > > Every time we try to hire a DSP programmer or coder or whatever you > want to call him/her, they always want to just consult. Hard to find > one that wants to come to work every day. > > But then, I find it hard to hire anyone technical these days, period. > They're evidently already working, retired, or so young all the do is > play with their wee's (oui's ?) all day.I dunno -- I've been in a number of places in the Portland Oregon area, and they're full of hardworking folks. Some young, some old. Maybe you need to hire some of the young ones, and train them up -- starting with the fact that, since you're delivering money to them, they're expected to deliver value in return. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" was written for you. See details at http://www.wescottdesign.com/actfes/actfes.html
Hiring a DSP programmer
Started by ●July 2, 2011
Reply by ●July 2, 20112011-07-02
Reply by ●July 2, 20112011-07-02
On Sat, 02 Jul 2011 16:45:08 -0700, Tim Wescott <tim@seemywebsite.com> wrote:>On 07/02/2011 03:44 PM, boB wrote: >> On Sat, 2 Jul 2011 12:24:56 -0700 (PDT), Jerry Avins<jya@ieee.org> >> wrote: >> >>> On Saturday, July 2, 2011 3:16:19 PM UTC-4, tinkerz wrote: >>>> I am after a control system based on frequencies, I need to gather >>>> information on the time and frequency domain. >>>> >>>> The data is fast and noisy, I made an assumption the fourier and laplace >>>> will be the place to start. >>> >>> You need to lay aside your preconceptions and find a good consultant. As a former consultant, I felt it was my duty to give the clients what they needed. That was often at odds with what they wanted. Tact usually won the day. >>> >>> Jerry >> >> >> Every time we try to hire a DSP programmer or coder or whatever you >> want to call him/her, they always want to just consult. Hard to find >> one that wants to come to work every day. >> >> But then, I find it hard to hire anyone technical these days, period. >> They're evidently already working, retired, or so young all the do is >> play with their wee's (oui's ?) all day. > >I dunno -- I've been in a number of places in the Portland Oregon area, >and they're full of hardworking folks. Some young, some old. > >Maybe you need to hire some of the young ones, and train them up -- >starting with the fact that, since you're delivering money to them, >they're expected to deliver value in return.Yes, that value part is what you never know until you hire them a while. We DID just hire a younger newbie that seems VERY good, knows his math and is interested in what we do. Also decent electronics designer. We got very lucky with him. I find the person hired must be interested and have a passion for what they do. The ones that just want a pay-check very rarely work out so well. I know good engineers are out there, but rarer these days... Just don't know which ones are THE good ones until on the job a while, except for those very few (like our new hire) that just scream out, "I'm the one" ! Portland has some good talent though, I'm sure. Look at all the tech companies there ! We should be on par up here near Seattle, but I'm not so sure. boB
Reply by ●July 2, 20112011-07-02
On Jul 3, 3:24=A0am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:> tinkerz wrote: > > Hi All > > > Where is the best place to hire a programmer for a DSP/Laplace control > > system > > > Thanks > > Tinkerz > > Programmer is a tool. What you are looking for is probably a coder. To > get a coder, go to a forest and find some orangutan.User interface is fulling functional I see Mr Impaler..
Reply by ●July 2, 20112011-07-02
On 07/02/2011 03:16 PM, tinkerz wrote:> I am after a control system based on frequencies, I need to gather > information on the time and frequency domain. > > The data is fast and noisy, I made an assumption the fourier and laplace > will be the place to start.If you're converting an analog signal to digital and performing signal processing in the discrete-time domain, then Laplace is the wrong analysis to apply in that domain; The z-transform is the equivalent to continuous-time's Laplace transform in the discrete-time domain. If you're processing the analog signal in the analog domain, then it isn't _digital_ signal processing. -- Randy Yates % "Watching all the days go by... Digital Signal Labs % Who are you and who am I?" mailto://yates@ieee.org % 'Mission (A World Record)', http://www.digitalsignallabs.com % *A New World Record*, ELO
Reply by ●July 2, 20112011-07-02
On 07/02/2011 05:02 PM, Tim Wescott wrote:> [...] > Randy is probably forgetting that the z transform (which is what is > used to analyze sampled-time systems) can be derived from the > Laplace transform.There may be some convoluted way to do that, but I don't think it's going to be of any utility. None of Lyons, Proakis, Mitra, or Oppenheim discuss deriving it from the Laplace transform. Of course it has the similar function as the Laplace transform in the digital domain, and as the Laplace transform is derived from the continuous-time Fourier transform, the z-transform is derived from the discrete time Fourier transform. However it isn't "derived" from the Laplace transform in any way I've seen or can imagine to be useful.> But normally when you talk about sampled-time control you talk about > the z domain, unless you're specifically modeling the plant, or are > trying to take an old continuous-time controller and recast it into > a sampled-time one.The plant can be modeled in the z-domain too. Sure, it may be analog in its original form, but if your controller or compensator is digital, some assumptions about maximum bandwidth must be made and the conversion to digital must be done somewhere. -- Randy Yates % "Watching all the days go by... Digital Signal Labs % Who are you and who am I?" mailto://yates@ieee.org % 'Mission (A World Record)', http://www.digitalsignallabs.com % *A New World Record*, ELO
Reply by ●July 3, 20112011-07-03
HardySpicer wrote:> On Jul 3, 3:24 am, Vladimir Vassilevsky <nos...@nowhere.com> wrote: > >>tinkerz wrote: >> >>>Hi All >> >>>Where is the best place to hire a programmer for a DSP/Laplace control >>>system >> >>>Thanks >>>Tinkerz >> >>Programmer is a tool. What you are looking for is probably a coder. To >>get a coder, go to a forest and find some orangutan. > > > User interface is fulling functional I see Mr Impaler..The orangutan Ph.D. ?
Reply by ●July 3, 20112011-07-03
"Randy Yates" wrote in message news:OMKdnZYsYaV0TpLTnZ2dnUVZ_tGdnZ2d@supernews.com... On 07/02/2011 05:02 PM, Tim Wescott wrote:> [...] > Randy is probably forgetting that the z transform (which is what is > used to analyze sampled-time systems) can be derived from the > Laplace transform.There may be some convoluted way to do that, but I don't think it's going to be of any utility. None of Lyons, Proakis, Mitra, or Oppenheim discuss deriving it from the Laplace transform. Of course it has the similar function as the Laplace transform in the digital domain, and as the Laplace transform is derived from the continuous-time Fourier transform, the z-transform is derived from the discrete time Fourier transform. However it isn't "derived" from the Laplace transform in any way I've seen or can imagine to be useful. <pm> Wouldn't it be better to say that the Fourier Transform is a special case of the Laplace Transform? </pm>> But normally when you talk about sampled-time control you talk about > the z domain, unless you're specifically modeling the plant, or are > trying to take an old continuous-time controller and recast it into > a sampled-time one.The plant can be modeled in the z-domain too. Sure, it may be analog in its original form, but if your controller or compensator is digital, some assumptions about maximum bandwidth must be made and the conversion to digital must be done somewhere. -- Randy Yates % "Watching all the days go by... Digital Signal Labs % Who are you and who am I?" mailto://yates@ieee.org % 'Mission (A World Record)', http://www.digitalsignallabs.com % *A New World Record*, ELO
Reply by ●July 3, 20112011-07-03
On Sat, 02 Jul 2011 07:22:37 -0500, "tinkerz" <nwoodhamuk@n_o_s_p_a_m.googlemail.com> wrote:>Hi All > >Where is the best place to hire a programmer for a DSP/Laplace control >system > >Thanks > >TinkerzHello Tinkerz, What is your company's name? What is your name? Is your company's headquarters in the United States? [-Rick-]
Reply by ●July 3, 20112011-07-03
On 07/02/2011 07:28 PM, Randy Yates wrote:> On 07/02/2011 05:02 PM, Tim Wescott wrote: >> [...] >> Randy is probably forgetting that the z transform (which is what is >> used to analyze sampled-time systems) can be derived from the >> Laplace transform. > > There may be some convoluted way to do that, but I don't think it's > going to be of any utility. None of Lyons, Proakis, Mitra, or > Oppenheim discuss deriving it from the Laplace transform. > > Of course it has the similar function as the Laplace transform in the > digital domain, and as the Laplace transform is derived from the > continuous-time Fourier transform, the z-transform is derived from the > discrete time Fourier transform. However it isn't "derived" from the > Laplace transform in any way I've seen or can imagine to be useful.Houpis & Lamont, "Digital Control". Shows how, if you model the sampling process as multiplying by a chain of impulses then the z transform just falls out of the Laplace transform. IIRC, Oppenheim, Wilsky & Young, "Signals and Systems" models sampling the same way, and derives the Fourier series from the Fourier transform; I'd be surprised if they don't derive the z transform the same way, later in the book.>> But normally when you talk about sampled-time control you talk about >> the z domain, unless you're specifically modeling the plant, or are >> trying to take an old continuous-time controller and recast it into >> a sampled-time one. > > The plant can be modeled in the z-domain too. Sure, it may be analog > in its original form, but if your controller or compensator is > digital, some assumptions about maximum bandwidth must be made and the > conversion to digital must be done somewhere.If you start with a Laplace domain model of the plant and you model the sampling and the reconstruction right (even if your 'reconstruction' is just a 1st-order hold) then you don't have to make any assumptions about maximum bandwidth -- you can construct an exact model of the plant, as seen from the processor, in the z domain. If necessary, the model can be easily expanded to allow you to predict what the plant will be doing in between processor sampling intervals, although these days that's usually just an indication that you're sampling too slowly. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" was written for you. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●July 3, 20112011-07-03
>On Sat, 2 Jul 2011 12:24:56 -0700 (PDT), Jerry Avins <jya@ieee.org> >wrote: > >>On Saturday, July 2, 2011 3:16:19 PM UTC-4, tinkerz wrote: >>> I am after a control system based on frequencies, I need to gather >>> information on the time and frequency domain. >>> >>> The data is fast and noisy, I made an assumption the fourier andlaplace>>> will be the place to start. >> >>You need to lay aside your preconceptions and find a good consultant. Asa former consultant, I felt it was my duty to give the clients what they needed. That was often at odds with what they wanted. Tact usually won the day.>> >>Jerry > > >Every time we try to hire a DSP programmer or coder or whatever you >want to call him/her, they always want to just consult. Hard to find >one that wants to come to work every day. > >But then, I find it hard to hire anyone technical these days, period. >They're evidently already working, retired, or so young all the do is >play with their wee's (oui's ?) all day. > >boB > > > > > >I just want to consult thats all, on a small job, with someone that knowswhat to do.
