The State has appealed, pursuant to Sec. 547.200.1 RSMo 1986, from orders by two different trial judges sustaining each defendant’s motion to suppress evidence seized by the Fredericktown Police Department. Defendant Canton was charged with misdemeanor possession of a controlled substance — marihuana. Defendant Keer was charged with felony possession of a controlled substance — marihuana. We denied the state’s motion to consolidate the cases but we choose to dispose of both cases in a single opinion.
The evidence suppressed was the marihuana and paraphenalia for smoking it. The evidence was seized at a roadblock established by two Fredericktown police officers on Highway 72 in Fredericktown five *353hundred feet west of the city limits. The roadblock occurred at 8:00 p.m. on March 10, 1989. It was dark and the only illumination at the site was provided by halogen lights at a school several hundred feet distant. The officers parked two patrol cars on each shoulder of the road with the emergency lights flashing but with the headlights off. Each officer had a flashlight, was dressed in uniform, but was not wearing reflective clothing. No flares or signs of any type were placed on the road to warn approaching drivers of a roadblock ahead. The speed limit at the site of the roadblock was 55 m.p.h.
The plan for the roadblock, and its location, was made by one of the officers sometime between 2:00 p.m. and when it began. That officer was the assistant chief of police. The police force had six members total and only the two conducting the roadblock were on duty. The chief of police was not aware of the roadblock until the next morning. There were no written plans within the police department for the operation of roadblocks, their location, the procedures to be followed, or when they should be conducted. The location for this roadblock was determined by the originating officer on some non-defined general awareness that some arrests had occurred in the area. The roadblock was established to check “for vehicle defects and check for operator’s license of the drivers and, also, possibly for any intoxicated drivers.” An additional reason “was to search vehicles for drugs.” As soon as a vehicle was stopped the officers immediately took their flashlights and began “looking for things” in the vehicle — “contraband and weapons.”
The roadblock commenced at 8:00 p.m. The procedures were discussed five minutes before that by the originating officer with the other officer. Every westbound vehicle was to be stopped. Some cars were stopped prior to the automobile driven by Canton and occupied by Keer and one other person. Upon looking into the automobile with his flashlight the originating officer saw a pipe on the floor partially concealed by a passenger’s lower extremity and the seat and a marihuana “bud” in the ashtray. The car was ordered to pull over and the occupants were arrested. Additional marihuana was found after the arrest. Because of a shortage of personnel, a vehicle which came along after the arrests was waved through without stopping. After the arrests the defendants were taken to the police station and the roadblock was suspended for four hours. After processing defendants the two officers returned to the site with another officer who had come on duty and resumed the roadblock for an additional forty minutes.1
The burden of proof is upon the state to show by a preponderance of the evidence that the motion to suppress should be overruled. Sec. 542.296.6 RSMo 1986. In State v. Welch, 755 S.W.2d 624 (Mo.App.1988), the Western District dealt at length with the constitutionality of roadblocks or checkpoints designed to locate and arrest intoxicated drivers. The court held that such checkpoints were not per se unconstitutional as a violation of the Fourth Amendment to the U.S. Constitution and Missouri Constitution Art. I, Sec. 15, prohibiting illegal searches and seizures. It did assert, however, that such a roadblock would be unconstitutional if in the method of conducting it the intrusion upon the individual rights of the drivers out-balanced the legitimate interests of the state.
Welch then set forth a series of factual considerations demonstrating that the checkpoint involved was established and operated in a way so as to cause minimal intrusion upon the rights of the drivers stopped. The court focused on at least four basic considerations in upholding the checkpoint. First it was conducted pursuant to a designed plan based upon specific data concerning alcohol related accidents occurring in the area. Second, the procedures were in writing, the checkpoint was established by written order of a command officer, the operation of the checkpoint was supervised by high ranking officers, and full instruction including a copy of the or*354der was given to field personnel. Third, the checkpoint was established in a way to give maximum notice to the drivers of what was ahead and to provide for their safety. Finally, the delay to the drivers was minimal. In State v. Payne, 759 S.W.2d 252 (Mo.App.1988) this district followed the Welch case and in doing so emphasized the requirement that the vehicles stopped must be on some random basis and not in a selective manner. Welch was also followed by the Southern District in State v. Vanacker, 759 S.W.2d 391 (Mo.App.1988).
These three cases all dealt with checkpoints established to detect alcohol impaired drivers and Welch stressed the state’s legitimate interest in the interception and removal of such drivers from the highways. We are unaware of any case in this state which has approved such checkpoints for the purpose of checking operator licenses, checking on the physical condition of vehicles or to search for drugs. These may be by-products of a sobriety checkpoint but they have not been approved as a reason for establishing such checkpoints. By the officer’s testimony this roadblock was specifically setup for those purposes as well as to locate intoxicated drivers. This is made manifest by the officer’s testimony that immediately upon stopping the vehicle he would check the interior for contraband.
Additionally, the roadblock here bears little or no resemblance to those approved in Welch, Payne and Vanacker. It was conducted without the knowledge or approval of the chief of police of this small police force. This is not to say that every checkpoint must have the approval of the highest ranking officer in the department, but there should be evidence that such authority has been vested or delegated to the officer who establishes the location and procedure at the roadblock. There was no specific data upon which the officer relied in establishing the location. There were no written procedures, orders, or adequate training of the participating officers. There was an inadequate number of officers involved to properly conduct any checkpoint as it was necessary to cease operation of the checkpoint, at least temporarily, as soon as a violation was suspected. There is no indication that any attempt was made to obtain additional personnel from within the department, from the sheriff’s office, or from the highway patrol.
Safety to drivers was nonexistent. The checkpoint was located in a 55 m.p.h. speed zone, 500 feet inside the city limits of Fred-ericktown and involved only vehicles entering the town from an essentially rural area. No signs, flares or other warning was given that a roadblock existed, the purpose of such a roadblock or that vehicles were expected to stop. The lighting in the vicinity was inadequate and the officers wore no reflective clothing. A driver approaching the scene would have no idea what was occurring, what his response should be, or why he was being stopped.
The cases heretofore cited have upheld roadblocks or checkpoints to locate intoxicated drivers where careful planning exists and the police have taken pains to assure that the intrusion on drivers’ rights and safety ⅛ minimal. We are not prepared to relax the standards established in those cases. The roadblock here does not meet constitutional standards and the trial courts properly sustained the motions to suppress.
Orders suppressing evidence affirmed.
STEPHAN and SATZ, JJ., concur.. The evidence at the two separately conducted suppression hearings was not identical but the basic facts relating to the validity of the roadblocks were in evidence at both hearings.